Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fisheries and trade of species Listed on CITES appendix II, with a focus on seahorses Kuo, Ting-Chun 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_november_kuo_ting-chun.pdf [ 5.1MB ]
Metadata
JSON: 24-1.0355398.json
JSON-LD: 24-1.0355398-ld.json
RDF/XML (Pretty): 24-1.0355398-rdf.xml
RDF/JSON: 24-1.0355398-rdf.json
Turtle: 24-1.0355398-turtle.txt
N-Triples: 24-1.0355398-rdf-ntriples.txt
Original Record: 24-1.0355398-source.json
Full Text
24-1.0355398-fulltext.txt
Citation
24-1.0355398.ris

Full Text

FISHERIES AND TRADE OF SPECIES LISTED ON CITES APPENDIX II, WITH A FOCUS ON SEAHORSES by  Ting-Chun Kuo  B.S., National Taiwan University, 2010 MSc., National Taiwan University, 2012  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   August 2017  © Ting-Chun Kuo, 2017   ii Abstract International trade regulations are important tools to ensure sustainable use of wildlife, but their effects on marine fishes are rarely examined. In this thesis, I evaluated patterns and changes in international trade of marine fishes whose exports are regulated under the Convention on the International Trade in Endangered Species of Wild Flora and Fauna (CITES). I focused on the case study of seahorses (Hippocampus spp.), which were the first marine fishes with exports controlled under CITES (Appendix II) in 2002, but also investigate international trade for all animal species on Appendix II.  In my first two research chapters, I investigated how implementation of CITES at the national level affected seahorse fisheries and trades. My study focused on Thailand, the country that has long exported the most seahorses. In Chapter 2, I showed that fishers reported different catch rates over varying reporting time periods. This correlation could greatly influence catch estimates and have implications for management that draws on fishers’ knowledge. In Chapter 3, I found that seahorse catch and domestic trade did not change after the Thai government set an export quota for seahorses, even though formal export data showed notable declines in volume.   In my next two research chapters, I considered trade dynamics at the global scale. In Chapter 4, I found that once the CITES Appendix II listing came into effect, total weight of seahorses in documented global trade decreased, recorded seahorse exports became dominated by few countries, and prices for seahorses increased. In Chapter 5, I explored trade patterns for all CITES Appendix II animal species. I discovered that the USA appears to be the centre of reported wildlife trade but also that several Asian countries were emerging as important   iii importers. I also found that the trade in marine fishes and non-coral invertebrates involved more countries with more connections than other Appendix II taxa.  My thesis highlights the opportunities and limitations of using export controls in managing the trade in marine fishes. These findings have implications for the implementations of global agreement, wildlife trade monitoring and conservation effort prioritizing.    iv Lay Summary My thesis takes a pioneering look at regulation of the global trade in marine fishes, with a focus on an international agreement called CITES. I was curious about how its first global commitment to limit exports of any marine fishes – to levels that do not damage wild populations - actually played out. These first fishes were the iconic seahorses. I found that although reported seahorse exports declined globally with the new regulations, just as many seahorses were caught in the biggest source country, Thailand. For all animal species under CITES regulations, I found that the USA was the centre of global wildlife trade, although Asian countries were gaining ground. I also found that trade in marine fishes involved more countries than most other animals on Appendix II. My study indicates that global agreements can contribute to conservation action and my findings can help setting conservation priorities.   v Preface  This thesis represents my own work. One chapter in this thesis (Chapter 2) is in revision for publication in a peer-reviewed journal, one chapter (Chapter 4) is in review, and the other chapters (Chapter 3, and 5) are in preparation for submission. I was primarily responsible for all the work in this dissertation, with guidance from Dr. Amanda Vincent. All field research in this dissertation was approved by UBC’s Behavioural Research Ethics Board (H12-02731). I list the contributions of my co-authors to each chapter below.  A version of Chapter 2 has been submitted for peer review with the title “Reporting time period matters: quantifying catch rates and assessing recall bias from fisher interviews in Thailand.” Dr. Lindsay Aylesworth is a co-author of this paper. In this work, I conceptualized the research question, conducted data analyses, and contributed to manuscript writing. Dr. Aylesworth provided the data used in the analyses, examined our results in light of Thai governmental trawl survey data, and contributed to manuscript writing. Authorship order was decided as L. Aylesworth and T.-C. Kuo, based on our relative roles in manuscript writing.  A version of Chapter 3 is in preparation for submission with the title “Implementation of CITES listing stimulates changes in trade: the case of dried seahorse trade in Thailand.” The co-authors of the manuscript are Dr. Parichart Laksanawimol, Dr. Lindsay Aylesworth, Ms. Ratanavaree Phoonsawat, Ms. Yooee Getpetch, Dr. Sarah Foster, and Dr. Amanda Vincent. Dr. Laksanawimol, Ms. Phoonsawat, and Ms. Yooee are Thai colleagues who provided important support during my fieldwork in Thailand. I conducted the fieldwork, analyzed the data and wrote   vi the manuscript. Dr. Laksanawimol, a Thai colleague, helped me conduct the interviews with fishers and traders, shared her knowledge of the location of the respondents, interpreted Thai and English during the interviews, and also conducted some interviews with higher-level traders. Dr. Aylesworth provided her data of the seahorse catch in Thailand to supplement the data I collected, and provided edits to the manuscript. Ms. Phoonsawat and Ms. Getpetch, also Thai colleagues, provided suggestions on fisher interview locations, shared their knowledge of Thailand’s fisheries management policy, and provided feedback to the manuscript. Dr. Foster and Dr. Vincent initiated the research, facilitated the collaboration with Thai government, secured the funding of this research, provided guidance during the fieldwork and data analyses, and provided edits to the manuscript.  A version of Chapter 4 is in preparation for submission, titled “Assessing the impacts of CITES implementation on the international trade of marine species – a case study of seahorses.” My co-author, Dr. Amanda Vincent, initiated the idea of this research, provided funding, gave feedback on interpretation of the analytical results and helped in writing the manuscript.   A version of Chapter 5, titled “Investigating patterns and changes in the declared trade of CITES Appendix II animals across taxa”, in is preparation for submission. My co-author, Dr. Amanda Vincent, helped generate the idea, provided funding, offered feedback throughout the analyses and helped in writing of the manuscript.   vii Table of Contents  Abstract .......................................................................................................................................... ii Lay Summary ............................................................................................................................... iv Preface ............................................................................................................................................ v Table of Contents ........................................................................................................................ vii List of Tables .............................................................................................................................. xiv List of Figures ........................................................................................................................... xviii List of Symbols ......................................................................................................................... xxiii List of Abbreviations ............................................................................................................... xxiv Acknowledgements .................................................................................................................. xxvi Dedication ............................................................................................................................... xxviii Chapter 1: Introduction ............................................................................................................... 1 1.1 Rationale ........................................................................................................................... 1 1.2 Background ...................................................................................................................... 3 1.3 Case Study: CITES .......................................................................................................... 8 1.4 Case Study: Seahorses ..................................................................................................... 9 1.5 Case Study: Thailand .................................................................................................... 11 1.6 Context ............................................................................................................................ 12 1.7 Research Questions ........................................................................................................ 12   viii 1.8 Thesis Outline................................................................................................................. 13 Chapter 2: Self-reported catch rate decreases with reporting time period: a case study of seahorse fishery in Thailand ........................................................................................................ 1 2.1 Synopsis ............................................................................................................................ 1 2.2 Introduction...................................................................................................................... 1 2.3 Material and methods ...................................................................................................... 5 2.3.1 Data Collection ............................................................................................................ 6 2.3.2 Calculating annual catch rates per fisher ..................................................................... 8 2.3.3 Differences across reporting time periods: among different fishers, each reporting only one time period (Among Fishers) ................................................................................... 8 2.3.4 Differences across reporting time periods: across the estimates from an individual fisher who reported more than one time period (Within Fishers) ........................................... 9 2.3.5 Comparing catch rates with external datasets ........................................................... 10 2.3.6 Implications for other seahorse studies ..................................................................... 12 2.4 Results ............................................................................................................................. 13 2.4.1 Differences across reporting time periods: among fishers reporting only one time period (Among Fishers) ........................................................................................................ 13 2.4.2 Differences across reporting time periods: within individual fisher who reported more than one time period  (Within Fishers) ........................................................................ 14 2.4.3 Comparing catch rates with external datasets ........................................................... 14 2.4.4 Implications for other seahorse studies ..................................................................... 15 2.5 Discussions ...................................................................................................................... 15   ix Chapter 3: Implementation of CITES listing stimulates changes in trade: the case of dried seahorse trade in Thailand ........................................................................................................... 8 3.1 Synopsis ............................................................................................................................ 8 3.2 Introduction...................................................................................................................... 9 3.3 Methods .......................................................................................................................... 14 3.3.1 Trade surveys - data collection .................................................................................. 15 3.3.2 Trade surveys - data analysis .................................................................................... 16 3.3.2.1 Estimate economic scale...................................................................................... 17 3.3.2.2 Estimate income from dried seahorses ................................................................ 19 3.3.2.3 Perceived changes in trade volume and prices over time and with CITES Implementation ................................................................................................................. 19 3.3.2.4 Examine changes in international trade using external datasets ......................... 19 3.4 Results ............................................................................................................................. 20 3.4.1 Trade structure ........................................................................................................... 20 3.4.2 Trade volume in dried seahorses ............................................................................... 22 3.4.3 Price of seahorses traded in Thailand ........................................................................ 22 3.4.4 Economic scale of dried seahorse trade in Thailand ................................................. 23 3.4.5 Income from trading dried seahorses ........................................................................ 24 3.4.6 Perceived changes in trade volume and price ........................................................... 25 3.4.7 Comparisons of Thai seahorse exports among different data sources ...................... 26 3.4.7.1 Trade volumes ..................................................................................................... 26 3.4.7.2 Prices ................................................................................................................... 27 3.5 Discussion ....................................................................................................................... 27   x Chapter 4: Assessing the impacts of CITES implementation on the international trade of marine species – a case study of seahorses................................................................................ 44 4.1 Synopsis .......................................................................................................................... 44 4.2 Introduction.................................................................................................................... 45 4.3 Methods .......................................................................................................................... 49 4.3.1 Trade data .................................................................................................................. 49 4.3.2 Analysis ..................................................................................................................... 51 4.3.2.1 Changes in declared seahorse trade after CITES listing ..................................... 51 4.3.2.2 Linking country-level characteristics to changes in seahorse exports ................. 52 4.3.2.2.1 Why some seahorse range states were documented seahorse exporters, while others were not? ............................................................................................................ 53 4.3.2.2.2 Why do some source countries purportedly continue their exports after CITES implementation, while others reported stop or having significant decline in exporting? ..................................................................................................................... 54 4.3.2.2.3 What determines the reported trade volume between two countries before and after CITES implementation? ................................................................................ 56 4.4 Results ............................................................................................................................. 57 4.4.1 Changes in declared seahorse trade after CITES ...................................................... 57 4.4.2 Linking country-level characteristics to changes in seahorse exports ...................... 59 4.4.2.1 Why some seahorse range states were documented seahorse exporters, while others were not? ................................................................................................................ 59 4.4.2.2 Why do some countries purportedly continue their exports after CITES implementation, while others stop or having significant decline in exporting? ............... 59   xi 4.4.2.3 What determines the reported trade volume between two countries before and after CITES implementation? ........................................................................................... 60 4.5 Discussion ....................................................................................................................... 61 Chapter 5: Investigating patterns and changes in the declared trade of CITES Appendix II animals across taxa ..................................................................................................................... 79 5.1 Synopsis .......................................................................................................................... 79 5.2 Introduction.................................................................................................................... 80 5.3 Methods .......................................................................................................................... 85 5.3.1 Data ........................................................................................................................... 85 5.3.2 Analysis ..................................................................................................................... 86 5.3.2.1 Examine properties of the wildlife trade network and compare the networks among taxa ........................................................................................................................ 87 5.3.2.2 Key range states, exporting and importing countries .......................................... 88 5.3.2.3 Exploration of the discrepancies between range-states and exporters ................ 88 5.3.2.4 Exploration of the changes in the number of species a country exported/imported ....................................................................................................................................89 5.3.2.5 Exploration of the changes in countries’ connections for the trade in CITES Appendix II animals .......................................................................................................... 89 5.4 Results ............................................................................................................................. 91 5.4.1 Examine properties of the wildlife trade network and compare the networks among taxa ……………………………………………………………………………………..91 5.4.2 Key range state, exporting and importing ................................................................. 92   xii 5.4.3 Exploration of the discrepancies between range-states and exporters ...................... 92 5.4.4 Exploration of the changes in the number of species a country exported/imported . 93 5.4.5 Exploration of the changes in countries’ connections for the trade in CITES Appendix II animals .............................................................................................................. 94 5.5 Discussion ....................................................................................................................... 94 Chapter 6: Conclusion .............................................................................................................. 114 6.1 Overview ....................................................................................................................... 114 6.2 Research Findings ........................................................................................................ 115 6.2.1 Research Question 1: How to get the best quantitative information for data-limited species in fisheries and trade? ............................................................................................. 115 6.2.2 Research Question 2: Has CITES impacted the trade in marine fishes at the national level?    ................................................................................................................................ 116 6.2.3 Research Question 3: Has CITES impacted the trade in marine fishes at a global level?    ................................................................................................................................ 117 6.2.4 Research Question 4: Do marine species have different trade patterns compared to other CITES Appendix II animals? ..................................................................................... 119 6.3 Implications .................................................................................................................. 120 6.4 Limitations.................................................................................................................... 122 6.5 Future Directions ......................................................................................................... 123 6.5.1 Reducing market demand for conservation of seahorses caught incidentally ........ 123 6.5.2 Management of bycatch in trawl fisheries .............................................................. 124 6.5.3 Prioritizing conservation efforts among CITES listed species ................................ 125 6.5.4 Addressing illegal wildlife trade ............................................................................. 126   xiii 6.5.5 Enhancing the role of industry in fostering sustainable wildlife trade .................... 127 6.6 Final remarks ............................................................................................................... 127 Bibliography .............................................................................................................................. 129 Appendices ................................................................................................................................. 171 Appendix A Supporting material for Chapter 2 ................................................................ 171 Appendix B Supporting material for Chapter 3 ................................................................ 180 Appendix C Supporting material for Chapter 4 ................................................................ 186 Appendix D Supporting material for Chapter 5 ................................................................ 197    xiv List of Tables Table 2.1 The median and mean of annual catch per vessel scaled up from various reported catch rates for trawlers reporting one time period. Kruskal-Wallis tests showed the mean ranks of scaled catch estimates were significantly different (p<0.01) among different reported time periods. ............................................................................................................................................ 1 Table 2.3 Results of pairwise comparisons (Dunn test) for annual catch of trawlers reported by time period. Numbers show the z-test statistics with p-values in the brackets for each pair of time periods. Significant values (<0.05) are italicized............................................................................ 2 Table 2.4 The median and mean of annual catch per vessel scaled up from various reported catch rates for gillnet fishers reporting one time period. Kruskal-Wallis tests showed the mean ranks of scaled catch estimates were significantly different (p<0.01) among different reported time periods. The results of the pairwise comparisons (Dunn test) show the z-test statistics with p-value in brackets. ............................................................................................................................ 3 Table 2.5 Mean ratio of each pair of scaled annual catch by reporting time period for fishers who reported multiple time periods. The mean of all ratios are shown, with sample sizes (n) of how many fishers reported on that pair of time periods. P-values <0.05 (bootstrap test) are italicized, representing mean ratios significant different to 1. ........................................................................ 4 Table 3.1 Number of respondents interviewed during trade surveys in Thailand, categorized by occupation and location. ............................................................................................................... 33 Table 3.2 Trade volumes and number of traders of dried seahorses for each trade level in Thailand. Data obtained from external datasets or literature are in italic. We used reported mean catch from otter trawlers to representative fishers’ trade volume (level 1). Number of traders was estimated based on the mean trade volume and by two methods (See Appendix B) for sensitivity   xv analysis: (1) considering both reported purchase volume and sell volume (P+S), and (2) considering only the maximum trade volume (T)......................................................................... 34 Table 3.3 Mean purchasing and selling price of dried seahorses in each trade level (USD per gram) across three seahorse size categories (Small (s), Medium (M), Large (L)). Sample sizes (number of respondents) are shown next to the prices in brackets. .............................................. 36 Table 4.1 Results of testing the effects of CITES interventions using iterative segmented regressions (Vt = a0 + a1Yeart + a2Yeart ∙ I + a3I). We tested each year from 2003-2007 for Hong Kong’s and 2000-2007 for Taiwan’s data as the break in each model. Here only we show the results from the model with the highest r-square, with p-value of each coefficient estimate in the brackets. The results for other models (with break point as other years tested) were shown in Appendix B. HK: data from Hong Kong CSD (1998-2014); TW: data from Taiwan Customs (1983-2014)................................................................................................................................... 68 Table 4.2 Results of gravity model for Hong Kong dried seahorse imports (Hong Kong CSD data), Taiwan’s dried seahorse imports (Taiwan Customs), and global bilateral dried seahorse trade (CITES data). Hong Kong and Taiwan’s imports in pre-CITES implementation (pre 2005) and post-CITES implementation (2005-2014) periods were separated in order to compare to the CITES data (2005-2014). Sample size (n) including trade volume = zero (set as 10-10) for potential country pairs in trade (see Method). All variables are in logarithm space. ................... 69 Table 5.1 The basic properties of the trade network for all CITES Appendix II animals and seven major groups of taxa, for data (a) from 1991-2014 and (b) 2014 only. ...................................... 103 Table A.1 Department of Fisheries Governmental trawls from 2010, 2012 and 2013. .............. 172   xvi Table A.2 Mean and median of annual catch rates per trawler based on Department of Fisheries’ surveys over a three year period. The annual rate was based on the catch rates for 4-hours haul...................................................................................................................................................... 173 Table A.3 Comparison of annual catch rates from Thai Department of Fisheries research trawls and fisher interviews in this study. Both mean and median annual catch per trawler are presented...................................................................................................................................................... 174 Table A.4 Comparison of annual seahorse catch rates for Trat province from fisher interviews (this study) and port-sampling (Laksanawimol et al. 2013). ...................................................... 175 Table A.5 Summary table shows the numbers of seahorse catch rates reported in previously published studies by their reporting time periods. ...................................................................... 176 Table A.6 Previously published studies reporting seahorse catch per haul but annual estimates based on another time period. We calculated annual estimates from reported haul values to compare with those published based on other periods................................................................ 177 Table B.1 Results of testing the effects of CITES interventions on Hong Kong’s import volume, using iterative segmented regressions (Vt = a0 + a1Yeart + a2Yeart ∙ I + a3I). .................... 187 Table C.1 Results of testing the effects of CITES interventions on Taiwan’s import volume, using iterative segmented regressions (Vt = a0 + a1Yeart + a2Yeart ∙ I + a3I). .................... 188 Table C.2 Results of testing the effects of CITES interventions on the number of source countries of Hong Kong, using iterative segmented regressions (Vt = a0 + aYeart + a2Yeart ∙ I + a3I)...................................................................................................................................................... 190 Table C.3 Results of testing the effects of CITES interventions on the number of source countries of Taiwan, using iterative segmented regressions (Vt = a0 + a1Yeart + a2Yeart ∙ I + a3I). .. 191   xvii Table C.4 Results of testing the effects of CITES interventions on the evenness of supply to Hong Kong, using iterative segmented regressions (Vt = a0 + a1Yeart + a2Yeart ∙ I + a3I). 193 Table C.5 Results of testing the effects of CITES interventions on the evenness of supply to Taiwan, using iterative segmented regressions (Vt = a0 + a1Yeart + a2Yeart ∙ I + a3I). ...... 194 Table C.6 Results of testing the effects of CITES interventions on Hong Kong’s import prices, using iterative segmented regressions (Vt = a0 + a1Yeart + a2Yeart ∙ I + a3I). .................... 196 Table D.1 The ISO-2 code of countries and territories that were mentioned in the main text. .. 197 Table D.2 The list of CITES Appendix II marine species .......................................................... 199 Table D.3 The results of linear regression of species richness in exports versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24). ................................................ 201 Table D.4 The results of linear regression results of species richness in imports versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24). ................................. 202 Table D.5 The results of linear regression of out-degree versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24) ............................................................ 203 Table D.6 The results of linear regression of in-degree versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24). ........................................................... 204 Table D.7 The results of linear regression of the ranking of closeness centrality versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24). ................................. 205   xviii List of Figures Figure 2.1 Commercial and small-scale fisher interview locations in Thailand............................. 5 Figure 2.2 Comparison of scaled annual catch reported by time period among fishers reporting only one time period. Seahorse catches from (a) trawlers, (b) gillnets for shrimp, (c) gillnets for fish, (d) gillnets for crab, and (2) other gillnets are shown, with number of fishers (n) reported their catch on each time period. ...................................................................................................... 6 Figure 2.3 Pairwise-comparison of scaled annual catch reported by time period from (a) trawlers and (b) gillnet fishers reporting more than one time period. For each fisher, we calculated the ratio of the two scaled annual catches for each pair of the time periods. If the catch estimates from two time periods are similar, the ratio should be close to one (dash line). Abbreviations for each time slice are: H: haul, D: day, T: trip, M: month, and Y: year. H:D represents the ratio of annual catch estimate from per haul to estimate from per day. ...................................................... 7 Figure 3.1 Provinces in Thailand surveyed during the trade surveys in 2013-2014. .................... 37 Figure 3.2 Dried seahorse trade structure in Thailand. Arrows indicate the direction of trade flow, from oceans to export. ......................................................................................................... 38 Figure 3.3 Potential trade routes for dried seahorses in Thailand as deduced from trade interviews. The arrows show the direction of trade flows. Survey locations are indicated as solid circles, whereas open circles represent places identified by respondents. .................................... 39 Figure 3.4 Changes in (a-c) mean trade volume and (d-g) per-gram selling price of dried seahorse reported by traders in each trade level in Thailand. Data were log transformed. The changes in trade volume and prices are presented by linear regression lines (if significant). The three interventions are marked by dashed lines: (i) CITES listing, (ii) CITES implementation, (iii) Thailand voluntary export quota implemented. ............................................................................ 40   xix Figure 3.5 A comparison of international trade volume from Thailand using three official datasets (CITES, Hong Kong CSD, and Taiwan Customs). Records from importers were stacked into one bar for each year; Hong Kong Census and Statistics data were in dark grey and Taiwan customs data were in light grey. For the CITES data, trade volume from Thailand to Hong Kong (HK) and Taiwan (TW) were shown in the same colour code as the Customs data but with stripes (HK: dark grey; TW: light grey), and to other destinations were in black. The three interventions were marked by arrows: (i) CITES listing, (ii) CITES implementation, (iii) Thailand voluntary export quota implemented............................................................................................................. 42 Figure 3.6 Comparing prices that Hong Kong (HK) and Taiwan (TW) imported seahorses from Thailand. The three interventions are marked by arrows: (i) CITES listing, (ii) CITES implementation, (iii) Thailand voluntary export quota implemented. .......................................... 43 Figure 4.1 Changes in seahorse trade (in kg) over time. Import data from Hong Kong CSD and Taiwan Customs are compared to the reports from export countries (RE) and import countries (RI) in the CITES database for years after 2005. The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled. ................................................................................ 71 Figure 4.2 Number of source countries of dried seahorses for Hong Kong and Taiwan. Data from the CITES trade databases considered reports from both export and import countries (Exports from imports). The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled. .................................................................................................................................... 72 Figure 4.3 Eveness (Gini Index) of the supply of dried seahorses for Hong Kong, Taiwan (CSD/Customs data + CITES data) and global trade (CITES data). The higher Gini index indicates less even in the supply. The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled. .......................................................................................................... 73   xx Figure 4.4 Changes in imports from each source country for (a) Hong Kong and (b) Taiwan. To see the changes in the proportions of imports from minor source countries, we show the cumulative imports to 40% of total imports in (c) and (d), for Hong Kong and Taiwan respectively. The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled. .......................................................................................................................................... 74 Figure 4.5 Changes in proportions of dried seahorses imports to Hong Kong from each source country and re-exports to each destination country. (a)-(h) shows the relative imports and re-export for every two years, and (i) shows the absolute import and re-export weight overtime. .. 75 Figure 4.6 Changes in import prices of dried seahorses for (a) Hong Kong and (b) Taiwan. Prices for seahorses from each source country are presented in different colours. Lines show the mean price of all source countries for each year, with 95% confidence intervals in gray. .................... 76 Figure 4.7 Comparisons of dried seahorse source countries (blue, n=32) and non-source countries (black, n=68) on the space of the first two principle components. Variables used in the PCA included demersal fish catch (Demersal catch), general trade value with China (Trade w. China), per capita GDP (GDP p.c.), marine fisheries employment, and distance to China (Distance). .... 77 Figure 4.8 Comparisons of seahorse source countries that stop exporting seahorses after the CITES listing (blue) and the countries that continue their exports (black). Variables used in the PCA included demersal fish catch (Dcatch), per capita GDP of each source country (GDP), the average seahorse trade volume in the pre-CITES period (preCITEStrade), marine fisheries employment (Employment), subsides for fisheries capacity building (sub.capacity) and beneficial fisheries subsides (sub. Beneficial). .............................................................................................. 78 Figure 5.1 Trade networks in 2014 for (a) all CITES Appendix II animals, (b) corals, (c) non-coral marine invertebrates and fishes, (d) mammals, (e) birds, (f) reptiles, (g) amphibians, and (h)   xxi insects. For (a), only countries with degree higher than average (n=51) were labeled. Line width represents the number of species in trade (in log scale), and line color corresponds to the region of the source country: Asia: red; Western & Southern Africa: brown; Northern & Eastern Africa: orange; Australia and Oceania: blue; America: purple; Europe: green. ..................................... 105 Figure 5.2 The number of CITES Appendix II animal species inhabiting each country’s terrestrial jurisdictions and/or EEZ. ............................................................................................ 107 Figure 5.3 The number of CITES Appendix II animal species reported being (a) exported from or (b) imported to each country from 1991-2014. ........................................................................... 108 Figure 5.4 The top 20 range states that exported the highest percentage of CITES Appendix II animal species that are native to them from 1991-2014. Countries are coded in official two-letter country codes (ISO2). The same information is displayed in map format in the inset. .............. 109 Figure 5.5 The top 20 countries exporting species that are not native to them from 1991-2014. Countries are coded in official two-letter country codes (ISO2). The same information is displayed in map format in the inset. .......................................................................................... 110 Figure 5.6 The percentage of all species (a) exported and (b) imported by a country over the total number of species in trade in each year, from 1991-2014. We show the top 10 countries in terms of exporting/importing the highest number of species in 2014, in descending order, and trace how the relative species diversity in trade for these 10 countries changes over time. Countries are coded in official two-letter country codes (ISO2). ..................................................................... 111 Figure 5.7 Changes in (a) out-degree and (b) in-degree for the 10 countries found to have the highest out-degree/in-degree in 2014. Countries are listed in descending order of importance along the y-axis. .......................................................................................................................... 112   xxii Figure 5.8 Changes in the ranking of centrality from 1991 to 2014 for the 10 countries ranked as having the highest centrality in the CITES Appendix II animal trade in 2014. Countries are ranked along the y-axis in descending order............................................................................... 113 Figure A.1 Department of Fisheries research trawl locations in the Andaman Sea and Gulf of Thailand. ..................................................................................................................................... 178 Figure A.2 Comparison of annual catch rates scaled from multiple time periods among fishers in different regions in Thailand. Seahorse catch from trawlers and gillnets are shown in (a) and (b) respectively. Southern Gulf of Thailand (S GoT) includes surveyed provinces Chumphon, Surat Thani, and Nakon Si Thammarat, Central and East Gulf of Thailand (C&E GoT) includes Samut Sakhon and Trat, and Andaman Coast (Andaman) includes Krabi, Trang, Phang-nga, Phuket, and Satun. .................................................................................................................................... 179 Figure B.1 Schematic illustrating the top-down calculations for number of traders in each trade level. ............................................................................................................................................ 181    xxiii List of Symbols  Yt – Reported trade volume or price at year t a0 – Initial value of an international trade variable (volume, price, number of countries, or Gini Index) at year 0 a1 - The change in an international trade variable with time a2 - The changes in slope of an international trade variable after considering the data after an intervention a3 - The level change in an international trade variable right after an intervention Dcatchit – Demersal fish catch of country i at year t Distij – Geographical distance between country i and j  GDPPCit – GDP per capita of country i at year t Fpopit – Number of fishers in the marine sector of country i at year t I – Intervention dummy variable  xijt – Trade volume between country i and j at year t β0 – Initial trade volume at year 0 β1 - The mean change in the trade volume/price in Thailand with time β2 – The level change in the mean volume/price in Thailand right after an intervention β3– The changes in slope of trade volume/price in Thailand versus year, after considering the data af   xxiv List of Abbreviations  AFCD – Hong Kong Agriculture, Fisheries, and Conservation Department CBD – Convention on Biological Diversity  CITES – Convention on International Trade in Endangered Species of Wild Fauna and Flora  CMS – Convention on Migratory Species  CPI – Relative consumer price index CPUE – Catch per unit effort  DMCR – Thailand Department of Marine and Coastal Resources  DoF – Thailand Department of Fisheries  EFI – Exports from imports EU – European Union  FAO – Food and Agriculture Organization of the United Nations HKD – Hong Kong dollar CSD – Hong Kong Census and Statistics Department GDP – Gross Domestic Product ICCAT – The International Commission for the Conservation of Atlantic Tunas IUCN – International Union for the Conservation of Nature  IUU – Illegal, unregulated, and unreported fishing  MPA – Marine protected area  NDF – Non detriment finding  PCA – Principal component analysis RFMO – Regional fisheries management organization    xxv RST – Review of significant trade  SD – Standard deviation TCM – Traditional Chinese medicine UBC – University of British Columbia UNEP-WCMC – The United Nations Environment Programme’s World Conservation Monitoring Centre USD – US dollar   xxvi Acknowledgements First and foremost, I would like to express my sincere gratitude to my thesis supervisor, Dr. Amanda Vincent. Thank you for guiding me get through this tough journey, inspiring me to think bolder, and being my role model as an independent professional woman working in conservation. I owe thanks to my committee members, Dr. William Cheung, Dr. Sumeet Gulati, and Dr. Tony Pitcher, for their support and encouragement during my PhD. I would also like to extend my thanks to: the UBC graduate studies, the Ministry of Education Taiwan, the Forestry Bureau of Taiwan, the UBC Ocean Leaders Training Program, the Hong Kong-Canada Business Association, the Thailand Department of Fisheries, the Hong Kong Chinese Medicine Merchants Association, People’s Trust for Endangered Species, Ocean Park Conservation Foundation Hong Kong, and Guylian Chocolates, for their financial and in-kind support. This thesis would not be possible without you.  I am thankful for all the support and friendship I gained during my fieldwork in Thailand, Hong Kong, and Taiwan. I sincerely thank P’Kiew, P’Jaeb, P’Yoo-ee, Pias, and Ake for their support and accompany when I was in Thailand. I also thank all the people who guided me and shared in my adventures when I was in Hong Kong and Taiwan: Mr. Tsang, Vincy, Sum, Wade, Mr. Chen, Mr. Ma, Mr. Kuo, and Po-Hsun. Most importantly, I really appreciate all of my respondents for their generosity in sharing their knowledge and life stories with me.   Thank you Dr. Sarah Foster for mentoring me throughout my dissertation and helping me to become a better scholar. Dr. Lindsay Aylesworth, I cannot express how lucky I am to have you as my friend and mentor. I cherish all the memories, and the laughter and tears we shared   xxvii together. Tanvi Vaidyanathan, I would never have made it without you. You were always there with me whenever I was up or down. I want to thank the students, post-docs, and staff at Project Seahorse: Danika, Jenny, Julia, Kyle, Ally, Clayton, Xiong, Emilie, Iwao, Gina, Scott, Riley, Lily, Tyler, Chai, and Tse-Lynn. I am so lucky to have this wonderful team always supporting me. I thank my dearest friends from the IOF: Sulan, Ravi, Mariana, Sahir, Xueyin, Shannon, Anna, Catarina, Roberto, Nicholas, Melanie, Madeline, Tayler, Travis, Jeff, Colette, and Andres. You made my life in Vancouver so much better.   Thank you my fellow Taiwanese graduate students in Vancouver: Sue-Jin, Yi-Chen, Jessie, Lindsay, Lavino, Rae, Justin, Amy, Linda, Santina, Stone, Jason, Owen, and Liz. You made Vancouver my home away from home. I also want to thank Yi-Chun, Kerker, Oscar, and Bug, for walking together with me through this PhD journey. I am truly grateful to have peers like you to chat about Ecology and so many other random things. Thank you Crystal and xmallwolf for running Ocean Says with me, it has been a great pleasure working with you. I thank my mentor, Dr. Chih-hao Hsieh, for leading and guiding me through the path of being a great scientist.  Special thanks are owed to my parents, who supported me in everyway they could. Thank you for encouraging me to follow my passion – even if you were always worried, you never wanted to discourage me. Thank you to my sister, for always cheering me up and taking care my parents while I was away. Thank you to my partner, Kai, for being an incrediblely thoughtful, patient, and supportive person in my life.    xxviii Dedication  This thesis is dedicated to my parents, Tzu-Chiang Kuo and Hsiao-Fen Hua. I also dedicate this thesis to those who dreamed of being scientists but could not. I cherish what I have and will work hard for you. Finally, I dedicate this work to all the animals and plants in this planet that are suffering because of human greed.  1 Chapter 1: Introduction  1.1 Rationale The impacts of international trade regulation for marine fishes are strikingly under-studied, despite the fact that many marine fish populations have been hugely depleted by intensive exploitation for trade. Although in recent decades, marine fishes have gradually been recognized as wildlife (not just food or commodities), they still received relatively less conservation attention than other vertebrates. Marine fishes are actually the dominant group of animals in international wildlife trade, in terms of products, volume, and value (Broad et al. 2001). The huge demand for wild-capture marine fishes has resulted in over-exploitation and habitat destruction by harmful extractive practices (Pauly et al. 2002, Pusceddu et al. 2014). Over-exploitation has led to more than 31% of the assessed marine fish stocks dropping below biologically unsustainable levels, and the proportion of the stocks over-exploited is still increasing (FAO 2016). In many regions, the catch of marine species is mainly destined for global markets rather than to the local populations, causing severe conservation and food security problems (Brashares et al. 2004, Swartz et al. 2010). However, despite the frequent and large-scale exchanges of marine fishes among countries, trade regulations for those species have mainly been applied at only national and regional levels, with the trade of few species being regulated globally (Vincent et al. 2013). While the lessons learned from terrestrial species indicate that effective trade regulation with international collaboration is essential in controlling the supply of threatened wildlife (Reeve 2006, Robinson et al. 2015), we have a poor understanding on how such regulations would affect the trade in marine species.    2 My dissertation examines the fisheries and trade of marine fishes under the wildlife trade management scheme, at a national and global level. I use the oldest and biggest environmental agreement on trade, the Convention on International Trade of Endangered Species of Flora and Fauna (CITES), as a case study. In the first three chapters, I focus on the fisheries and trade of seahorses (Hippocampus spp.). Then I expand my scope to the global trade patterns of all animal species under CITES Appendix II, with a focus on marine species.  Seahorses are the first fully marine fishes listed in CITES Appendix II since its inception. The listing of seahorses in CITES means that export of these 41 species has to be restricted to levels that are not detrimental to the wild population, and that specimens must be legally sourced. Seahorses provide an invaluable example for an investigation into CITES effects because of the global nature of the trade in seahorses (about 80 countries) and the large volumes of these fishes in trade (tens of millions of individuals p.a.). Crucially too, pre-CITES trade data exists for seahorses, thus providing a baseline from which to evaluate the treaty’s impacts on declared trade.   In Chapter 2 and 3, I examine fisheries and trade in seahorses at a national level, with the latter considered in the context of a CITES listing. I then report on the first study investigating changes in the trade of CITES-listed marine fishes at a global level (Chapter 4). Finally, I probe trade of all animal species under CITES Appendix II, seeking patterns while comparing marine fishes and all other taxa (Chapter 5). My findings will help support the implementation of trade management for other marine fishes, with the goal of achieving sustainable trade.    3  1.2 Background Regulating fisheries and trades in wild-caught species is essential to ensure the sustainable use of wildlife. Monetary and consumptive value of wildlife has driven exploitation of species beyond just supporting livelihoods, and has increased the pressure on wildlife survival. With growing human populations and greater globalization, wildlife trade has become ever more prevalent and lucrative (Hilton-Taylor et al. 2009, Bush et al. 2014). Wildlife trade includes all exchange of wild animals and plants, involving more than 40,000 plants and animal species in a mixture of live individuals and diverse products (Oldfield 2003, Nijman 2010). Although the volume and value of wildlife trade is hard to estimate given its often illicit nature, a study in 2008 estimated that global wildlife trade was worth around US$333 billion annually (Engler 2008). The trade in wildlife is not only for subsistence purposes (i.e. food, clothes, and medicines), but also for use as luxury goods, animal collections, pets, medicines, exotic foods and more. However, poorly regulated wildlife trade may result in detrimental consequences, such as population decline, invasive species, and the spread of diseases (Smith et al. 2009). In particular, over-exploitation is the biggest threat for many species (Maxwell et al. 2016, Harrison et al. 2016).   To ensure the sustainability of wild animals and plants, national governments have adopted both extraction controls and trade regulations. Some legal instruments directly restrict aspects of wildlife extraction (e.g. hunting and fishing), such as closed seasons, closed areas, size limits and catch quotas. By setting a fixed standard, these regulations provide clear guideline for law enforcement. Because the incentive for extracting wildlife is the economic gain from selling the species, trade regulations can be effective in controlling the supply. Commonly used trade   4 regulations include trade bans, trade quotas, tax, and tariffs (Roe et al. 2002). Because some extractive restrictions are easier to enforce than some trade restrictions, and vice versa, a combination of regulatory approaches often works best for wildlife conservation (Sterner 2003).   A key to successful wildlife managements is to understand the extraction and trade over time and their responses to management measures, however such monitoring is often insufficient. Monitoring is important to evaluate the management effectiveness, and provide information for adaptive management (Lyons et al. 2008). While monitoring can be costly and require professional capacity, efficient monitoring and evaluations for the management is sometimes overlooked or implemented poorly (Walsh and White 1999, Song 2008). National government needs to continue gathering information to improve their management.  Despite national efforts to manage wildlife trade, international demand poses substantial challenges for conservation by providing additional incentives to extract wildlife and intensifying the exploitation pressures (Oldfield 2003). Moreover, it is well documented that industrialized countries consume a greater proportion of traded environmental resources than developing countries (Watkins & Fowler 2002), indicating uneven environmental pressures on different regions. Once the supply from a source country becomes depleted, the international markets will turn to new sources, creating a story of serial exploitation of populations across regions or for similar species (Scales et al. 2007, Anderson et al. 2011). Furthermore, because legislation varies among countries, illegal exports from the countries with stricter regulations may hide behind products from countries with more relaxed regulations, thereby making illegal trade more difficult to detect (Bennett 2015). In addition to placing extra pressure on wild   5 populations, international trade can facilitate non-native species being introduced into new regions. These introduced species may potentially compete for resources and increase predation of native species, as well as the introduction of exotic pathogens (Smith et al. 2009).  As international trade is being accelerated by globalization, collaboration among countries is critical to prevent further over-exploitation. National governments may employ export/import regulations to ensure that sourcing of wildlife follows environmental, human rights, or/and health standards. Some of the unilateral import regulations, such as the well-known Tuna-Dolphin case from the United States, which bans the imports of tuna products from fishing practices that are harmful to dolphins, have resulted in substantial changes in trades and fisheries in source countries (Leroy et al. 2016). However, the large-scale and complicated international wildlife trade requires global collaboration to secure wild conservation of populations.  Elements of this collaboration will include coherent actions, experience sharing, capacity support, and commitments from both import and source countries. The Convention on International Trade of Endangered Species of Flora and Fauna (CITES) was therefore established in 1975, currently with 182 countries as members (plus the European Union). CITES places species on three Appendices according to the extent of the threats they face. The trade in species listed in the Appendix I is prohibited, and CITES member countries need to ensure that the trade in the Appendix II species to be not detrimental to the wild populations and is legally sourced.   In spite of being the most heavily exploited animal species, the use of marine fishes has been mostly regulated at the national and regional level, with global trade regulations applied to only few species. In 2014, more than 81 million tonnes of marine fishes were captured for human   6 consumption and other uses, with about 78% of those products exported (FAO 2016). Among the hundreds of millions of plants and animals that are traded every year, fisheries products are estimated to compose one-quarter of the total traded wildlife in terms of value (Engler 2008). National governments’ restrictions are mainly on entry to the fisheries or on the catch, and only few cases include regulations on trade (e.g., restrictions on selling catch from recreational fisheries) (Arlinghaus and Cooke 2009). Regional fisheries management organizations (RFMOs) are structured to manage the stocks in open oceans or/and migratory species. Several RFMOs (e.g., ICCAT) have recognized the potential of trade measures in increasing compliances to fisheries regulations (Tyler 2006), and trade sanctions issued by RFMOs on some countries have resulted in decreasing imports of selected species from those countries (Hosch 2016). However, few commercially important species are protected by RFMOs, and national waters, which are often the most exploited marine ecosystems, are not within RFMOs’ jurisdiction. In addition, RFMOs are often criticized for lacking enforcement capacity and being ineffective in reducing the pressures on bycatch species (Gilman et al. 2014).   Multilateral environmental agreements have gradually – although only recently - expanded their scope to include marine species because of the severe threats they faced or the large scale and complicated nature of the trade in those species. For example, the Aichi Targets of the Convention on Biological Diversity (CBD) require each country to include 10% of their waters in MPAs. The Conservation of Migratory Species of Wild Animals (CMS) has created restrictions on several commercially important marine species, such as sharks, manta rays, and sawfishes since 2002 (most since 2014). CITES has slowly been adding marine species to Appendix II since 2002 (Vincent et al. 2013), including 13 sharks and rays species listed in   7 Appendix II in 2016. There is an urgent need to examine the effectiveness of those treaties and how countries implement these multilateral agreements.  The long history of seeking a balance between international trade and conservation has resulted in some encouraging progress, primarily with terrestrial animals. Most cases of species population increases after CITES interventions are for Appendix I species, with all international trade banned (e.g., citron-crested cockatoo Cacatua sulphurea citrinocristata and giant otter Pteronura brasiliensis, Cahill et al., 2006; Recharte Uscamaita and Bodmer, 2010). The most positive example of Appendix II species, which legal trade is allowed under regulations, comes from crocodilians. With intensive hunting pressure, all crocodilian species were listed on CITES Appendix I in 1975. Fortunately, the trade has largely become sustainable, to the point where 12 of the 23 crocodilian species originally on Appendix I have been down-listed to Appendix II (although some still have populations on Appendix I). Indeed these twelve crocodilian species – all heavily traded – are now considered of Least Concern by IUCN Red List, even though they are still retained on Appendix II. The success is mainly due to the development of captive-breeding and ranch programs, and a huge amount of investment in managing sustainable hunting for wild crocodiles (Hutton and Webb 2002). However, not all species under trade regulations are as lucky as the crocodiles. Trade regulations, as with many other conservation/management measures, can fail easily because of inadequate management strategy, insufficient enforcement, difficulties in captive breeding, and so on (Smith and Benítez-Díaz 2011, Sand 2013, Wiersema 2013). Understanding the success and failure in each case will provide insight into future trade management, especially in the little evaluated cases for marine species.    8 1.3 Case Study: CITES I used the case study of CITES to examine how international export regulations may affect trade in marine fishes.  CITES is one of the biggest and oldest Conventions for sustainable trade. Currently, 182 nations are members of this Convention. The teeth of CITES are its three Appendices. All species in Appendix I are at high risk of extinction, such that the trade of those species, if wild-sourced, is basically prohibited. Species listed in Appendix II are not necessarily threatened with extinction, but their international trade must be controlled in order to avoid being detrimental to the wild populations. All specimens being traded have to be legally sourced and the transportation of live animals need to follow welfare guidelines. Appendix III species are protected in at least one country, which has asked other parties to support the regulation. When a species on Appendix III is exported, a permit from the source country is required.  Compared to other multilateral environmental agreements, the enforcement processes of CITES should provide the treaty strength to achieve sustainable trade. CITES requires Parties that intend to export any Appendix II species to make Non-Detriment Findings (NDFs), certifying that the trade is sustainable. To produce positive NDFs, countries need substantial information on the trade and population status of the species, and assessments on the risk of the exports (in addition to domestic trade) to their population survival. Where there is doubt about the actions of Parties to fulfill CITES requirements, they may be brought into CITES’ Review of Significant Trade (RST). If the Parties fail to address the concerns and recommendations made in RST, their trade may be suspended.    9  Marine fishes came under the purview of CITES only recently. Fisheries target and incidentally catch various animals including fishes, crabs, shrimps, corals, clams, and squids, and most of them end up in trade. However, until 2002, no fully marine fish had been listed in CITES since its inception. That was due to a persistent dispute about whether trade in commercial fishes should be regulated by RFMOs or the Food and Agriculture Organization (FAO) rather than CITES (Vincent et al. 2013). In 2004, CITES amended its listing criteria, adding footnotes to Annex 5 (Definitions, explanations and guidelines) specifically to explain the application of some criteria for commercially exploited aquatic species (CITES 2004). Since then, there has been a trend of proposing marine organisms in CITES (Cochrane 2015). Given that more species are being added to CITES, including at the Conference of Parties (CoP) in 2016, an understanding of trade in currently CITES-listed species will be valuable in guiding action for the more recently listed species, of greater commercial value.  1.4 Case Study: Seahorses Three of my thesis chapters focus on the trade of iconic marine fishes, i.e., the 41 seahorse species in the genus Hippocampus. Seahorses are suffering from an array of anthropogenic pressures, including direct extraction, incidental capture, and degradation or loss of their inshore coastal habitats. Their unique life history traits, such as monogamous mating patterns, small brood sizes and intensive male parental care – allied with patchy distribution, low mobility and small home range sizes – mean that those species may be more vulnerable to disturbances (Foster & Vincent 2004). Currently, one seahorse species (H. capensis) is classified as   10 Endangered and 11 seahorse species as Vulnerable in the IUCN Red List of Threatened Species, while the rest of the genus is Data Deficient.   Seahorses are caught in both directed and indiscriminate fisheries and then traded for traditional medicine, curios, and aquarium display. More than 37 million seahorses may be caught in non-selective gears in each year, primarily in trawls or gillnets (Lawson et al. 2017). Official trade records documented exports of 3.3 to 7.6 millions of seahorses legally per annum from 2004 to 2011 (Foster et al. 2016), but field surveys show that tens of millions end up in illegal, unreported, or unregulated trade (Vincent and Foster 2017). Such international trade involves three-quarters of the known species (31/41) and 87 countries (Foster et al, 2016). Among all countries, Thailand has apparently been the largest exporter of seahorses, responsible for at least 75% of the reported global export volumes each year from 2004-2011 (Foster et al. 2016). The vast majority of seahorses are traded dry, for use in traditional Chinese medicine (TCM) and as curios (Vincent et al, 1996).  In 2002, seahorses became the first fully marine fishes listed in CITES Appendix II since its inception, along with whale sharks (Rhinocodon typus) and basking sharks (Cetorhinus maximus). Seahorses are particularly useful for investigating the influence of CITES because of the availability of pre-CITES data, including trade surveys in source countries and Customs records in two major markets, which is very rare in other CITES-listed marine species. Two trade surveys of seahorses had been conducted in the early and late 1990s, before their CITES listing (Vincent 1996, Perry et al. 2010). In addition, pre-CITES official Customs records of dried seahorses are available for the two major markets, Taiwan and Hong-Kong SAR, with a   11 time series starting in 1982 and 1998, respectively, and continuing (independent of CITES data) to the current day.   The listing of seahorse species in the CITES Appendix II set precedent for other marine fishes. The first NDF framework for marine species was designed for seahorses (CITES 2013), and has provided a foundation for NDFs on sharks (Mundy-Taylor et al. 2014). Seahorses are also the first marine fishes to a Review of Significant Trade (RST: Vincent 2013). The RST of seahorses in Thailand, Vietnam, Guinea, and Senegal identified conservation concerns, and resulted in CITES imposing a suspension of the exports of H. kuda from Vietnam and H. algiricus from both Guinea and Senegal. In addition, Thailand suspended its own exports of all seahorse species when it became obvious that the country could not adequately implement the recommendations under RST (CITES 2016a).  1.5 Case Study: Thailand My first two chapters focus on the fisheries and trade of seahorses in Thailand, the country with the largest seahorse exports to date. Thailand contributes to a mean of 75% of global trade volumes of seahorses per year since 2005 to 2011 (Foster et al. 2016). Because of the large export volume, Thailand has been subject to CITES’ Review of Significant Trade and designated “urgent concern” by CITES Animal Committee (CITES 2012). Since Thailand was unable to scientifically justify if their exports were sustainable, CITES Animal Committee had made recommendations to Thailand to improve their NDF assessment (CITES 2012). In 2016, Thailand decided to suspend its seahorse exports before it can ensure the trade to be sustainable (CITES 2016a).   12 1.6 Context As a PhD student in Project Seahorse, I have a wonderful opportunity to conduct research on CITES and seahorse trade. Project Seahorse first uncovered the trade in seahorses and has led all seahorse conservation work in response, around the world. It acted as Chair of the seahorse working group for CITES and also prompted the CITES Review of Significant Trade to examine implementation of the Convention. In addition, Project Seahorse has developed collaborations with scientists from university and government in Thailand to help the country to meet CITES requirements. Our Thai collaborators include the Department of Fisheries (DoF), Department of Marine and Coastal Resources (DMCR), Department of National Parks (DNP), and Kasetsart University. Project Seahorse signed a formal MoU with Thai DoF to conduct research in Thailand, with endorsement from the CITES Secretariat in Geneva. The collaboration provides opportunity for me to collect data for my first two chapters to understand how the seahorse fisheries and trade in Thailand affected by the CITES listing. Further, Project Seahorse experience with and assessment of CITES trade databases allowed me to avoid some of the obvious pitfalls of these records.  1.7 Research Questions In this thesis, I focused on the following four questions to understand CITES impacts on the trade in marine fishes: (1) How to obtain the best quantitative information for data-limited species in fisheries and trade? (Chapter 2) (2) Has CITES impacted the trade in marine fishes at the national level? (Chapter 3) (3) Has CITES impacted the trade in marine fishes at a global level? (Chapter 4)   13 (4) Do marine species have different trade patterns compared to other CITES animals? (Chapter 5)  1.8 Thesis Outline This thesis has four research chapters, followed by a general discussion of the implications, strength and limitations of these studies. The research chapters (Chapter 2 to 5) will be in plural personal pronoun (“we”) since they are team-authored works, and the contributions of each co-author is listed in the Preface.  In Chapter 2 and 3, I investigate the fisheries and trade of seahorses in their biggest source country, Thailand. In Chapter 2, I explore a source of variation and potential recall bias in estimating annual seahorse catch based on fishers’ interviews. Given that they are predominantly a bycatch in non-selective gear, fisheries data on seahorses is rare. Fishers’ interviews provide baseline information for these data-limited fishes but the reliability of the interview data is affected by fishers’ recall biases. I here examine an important source of recall bias: the time period fishers use to report their catch rates and how it affects estimates of seahorse catch. In Chapter 3, I estimate the economic value of seahorse fisheries and exports for Thailand. Furthermore, I examine how the Thai government’s implementations of the CITES listing affected the seahorse catch and trade.   In Chapter 4, I probe whether and how international seahorse trade changed at the time of the CITES listing and implementation, using Customs records from two major markets, Hong Kong and Taiwan, which were initiated well before listing and continue today. I also incorporate   14 CITES trade data which record global seahorse trade, to explore how the characteristics of a country affect its seahorse exports and imports.  In Chapter 5, I expand my focus to all animal species listed in CITES Appendix II, probing different patterns in trade networks between marine species and other animals. I also identify the important range states, exporters, and importers, in terms of species richness in trade and connections with other countries, across all taxa. This work aims to help in prioritizing conservation efforts at a global scale for multiple species.  In the last Chapter, I discuss the major findings presented in this thesis, and the implications of those findings to wildlife trade management.   1 Chapter 2: Self-reported catch rate decreases with reporting time period: a case study of seahorse fishery in Thailand  2.1 Synopsis Catch rates reported by fishers are commonly used to understand the status of a fishery, but the reliability of fishers’ reported data is affected by how they retrieve such information. Using interview data from trawlers and gillnet fishers in Thailand, we investigated how the time period for which fishers reported their catch rates (e.g., per day or month) correlated with annual catch estimates. We examined such correlations by comparing (1) across single estimates reported by different fishers (Across Fishers), and (2) across different estimates reported by single fishers (Within Fishers). Across Fishers, we found that the annual catch estimates of fishers who reported on a shorter time period (haul, day) were significantly higher than those who reported on a longer time period (month, year). Within Fishers, where single fishers told us about catches on multiple different time periods, annual catch estimates were again higher on shorter time periods, in a form of recall bias. By comparing the reported catch rates with external datasets, we identified that the mean across all reports was the most similar to other data sources, better than using any one particular time period. Our research suggests that catch estimate is associated with reporting time period, which has strong implications for fisheries management.  2.2  Introduction In many situations, fisheries management relies on fisher knowledge to gather information (Neis et al. 1999). Through their experience interacting with the marine environment, fishers obtain an   2 understanding of catches, catch per unit effort, changes over time, key habitats, and historical catches (Neis et al. 1999, Lozano-Montes et al. 2008, Beaudreau and Levin 2014). However, the reliability of fisher knowledge is often questioned because it is not easily compared with scientific data and is difficult to verify or corroborate (Usher 2000, Byg et al. 2012). Despite these challenges, numerous studies show that fisher knowledge can nicely complement scientific information (Lozano-Montes et al. 2008, Taylor et al. 2011, Beaudreau and Levin 2014).  Fishers’ reports tend to vary greatly in ways that influence fisheries management, both because of differences among fishers or/and because of systemic errors (biases) (Johannes et al. 2000, Erisman et al. 2015). Variations among fishers’ reports have been documented by gender, age, geographical location and targeted species (Johannes et al. 2000, Erisman et al. 2015), while the exploitation rates of fishers also vary in space and time (McShane 1996, Barbeaux et al. 2014). In addition, fisher knowledge is influenced by his or her own awareness and may only relate to targeted species and identifying patterns that maximize capture success (Hamilton and Walter 1999). The perceptions and motivations of fishers can influence their behavior, awareness, and reporting of events (Carr and Heyman 2012, Panjarat and Bennett 2012, Bennett and Dearden 2014), which may lead to bias when sharing information for fisheries management.    Some variations among fishers, such as the frequency of encountering an event, may result in different cognitive processes when fishers report quantitative information (e.g., catch rates) (Beaman et al. 2005). For example, fishers who rarely catch a certain species may recall their catch using “episode enumeration”, recalling the event from long-term memory within a relevant time frame (Sudman and Bradburn 1974). In contrast, fishers who catch a species more   3 frequently may report their catch rates using a “rule-based” process (Burton and Blair 1991, Vaske et al. 2003). That is, they may recall their catch based on the time period they feel most relevant (e.g., 5 fishes per day) and then calculate the catch rates by considering the frequency of their fishing activity (e.g., 6 fishing days per week), to respond to a question with a given time frame.  The different cognitive processes suggest that recall biases may emerge if respondents report on time periods that are less relevant to their fishing activity (Tarrant et al. 1993, Golden et al. 2013). Recall bias, or people’s ability to access and retrieve information and experiences of the past, can greatly affect a respondent’s reporting of events (Denzin and Lincoln 2011). Within the broader literature of recall bias, patterns have emerged highlighting issues related to time period (Denzin and Lincoln 2011). Some studies have found that hunters over-estimate their harvests when reporting on a shorter time period, if the successful exploitation is a rare event (Golden et al. 2013). Variations in self-reported metrics linked to reporting time period have been found in some recreational fisheries such as with the number of days fished and fish consumption (Connelly and Brown 1995). In terms of fisheries management for data-limited species, we are often interested in a positivist approach (Roth and Mehta 2002), where we quantify the data reported by fishers as a means to understand the truth or objective reality of the fishery (O’Donnell et al. 2010, Shepperson et al. 2014). Key to our interpretation of such data is the understanding of any recall bias and its influence on data analysis and management decisions.  While there are numerous ways to address recall bias in fisher interviews (Denzin and Lincoln 2011, Carr and Heyman 2012), one method that deserves more scrutiny is triangulation.   4 Triangulation involves asking multiple questions that relate to the recall of an event (Denzin and Lincoln 2011). In fisheries, metric triangulations are used to cross-validate catch estimates reported by respondents (O’Donnell et al. 2012). For example, a fisher may be asked to report a catch estimate per day, week, month or year, as a way to triangulate the “true” catch rate for a given species. However, if the recall bias regarding to reporting time period exists, does triangulation really help to cross-validate the reported catch rates? Investigating such biases could have important implications for scaling catch rates to annual fishery estimates and thus providing guidance for data-limited fisheries (Johannes et al. 2000). However this potential source of bias has yet to be fully examined in a fisheries management context.   Data-limited fisheries are common throughout the world, particularly in a small-scale setting or when species are caught as bycatch (Patrick et al. 2009, Moore et al. 2013). One example of such fisheries is global seahorse fisheries, for which data are very sparse and patchy. Seahorses (Hippocampus spp.) are a genus of fishes often obtained as bycatch and then sold into international markets (Foster and Vincent 2004). Since there are growing markets for dried and live seahorses, and because seahorses have a unique shape, fishers often notice their occurrence in the catch. Several studies have investigated the use of fisher knowledge to provide baseline information on seahorse distribution, habitat, trade and catch rates (Baum et al. 2003a, Salin 2005, Giles et al. 2006, Perry et al. 2010). Only one study has investigated the potential for bias in fisher recall related to seahorses with a comparison of interviews and logbook data (O’Donnell et al. 2012). The results from this target fishery in the Philippines found that recall bias was evident as fishers perceived declines that did not match recent logbook entries (O’Donnell et al.   5 2012). For non-target seahorse fisheries, fisher knowledge, its use and potential for bias, has yet to be explored.    Our paper explores if annual catch estimates for non-target seahorse fisheries differ depending on fishers’ reporting time periods, and the potential influence of recall bias in reported catch rates. We examined two hypotheses regarding to reporting time periods: (1) annual seahorse catch estimate would vary among fishers according to the time periods they chose to use in reporting (day, week, fishing trip, month, year); (2) annual catch estimates would vary among different time periods reported by an individual fisher, indicating recall bias. We use a case study from commercial and small-scale fishers from Thailand reporting catch rates of a non-target fish genus, Hippocampus spp. Thailand is the world’s largest exporter of seahorses, the majority of which are obtained as bycatch in both commercial and small-scale fisheries (CITES 2013, Laksanawimol et al. 2013). Our analysis will enhance an understanding of the importance of reporting time periods and the potential source of recall bias, which has implications for fisheries management, especially for those data limited or those where fishers self-report.  2.3 Material and methods We elicited seahorse (Hippocampus spp.) catch rates in Thailand through semi-structured interviews of fishers in the commercial and small-scale sectors. Our paper focuses on the two dominant gear types we found during interviews –trawlers (commercial) and gillnets (small-scale). These gears also represent the dominant gear types in commercial and small-scale fisheries according to the Thailand Department of Fisheries. Thai fishers kept no log-books or landings data for seahorse catch. We therefore collected information through fisher interviews.    6  2.3.1 Data Collection To determine annual seahorse catch per unit effort from fishing gears in Thailand, we conducted semi-structured interviews from 2013-2014 with commercial and small-scale fishers. We interviewed a total of 238 fishers (86 commercial trawlers, 152 small-scale gillnets) in 11 provinces on both the Andaman (n=6) and Gulf (n=5) coasts of Thailand (Figure 2.1). We only interviewed fishers over the age of 18, who were actively fishing. Of the 238 fishers interviewed, eight were women. Of the 86 trawl fishers interviewed, only 4 fishers were unable to provide catch rate estimates. Of the 152 gillnet fishers, 111 reported using different types of gillnet to target different species (e.g., shrimps, crabs), and reported catch estimates for each gillnet type. We treated these responses independently and categorized them according to the type of gillnet, leading to 58 responses for shrimp gillnet, 82 for fish gillnet, 97 for crab gillnet, and 23 for others.   To select respondents, we targeted two sets of landing locations: fishing ports for commercial fishers and coastal villages for small-scale fishers. We determined the number of fishers to interview in each location as either 10% of the estimated total number of fishing boats landing catch at that location or the saturation method (Tobias 2010) given our need to maximize the number of landing locations visited. We defined saturation as the point where if the 6th, 7th, and 8th, interviews introduced no more new information on gear type, fishing grounds, or seahorse locations, then saturation had been achieved. We additionally confirmed saturation through our observations of gear type at each landing location. To understand the current situation of seahorse fishery (2013-2014), we excluded interviews from fishers who had retired from fishing   7 at the time the survey was conducted. All interviews followed UBC Human Ethics Protocols (H12-02731).  During the semi-structured interviews fishers were asked questions about their current seahorse catch and effort. Fishers were asked to identify the gears they use to capture seahorses, along with numbers and frequency of catch. Fishers reported their catch rates over the time period they preferred and we categorized their responses post-hoc. The time periods fishers used to report seahorse catch included per haul, day, fishing trip, week, month and year. Fishers might also report catch rates in more than one time period. For example, some fishers reported average seahorse catch per day and also per month. The reported catch rates would be scaled up to annual estimates per vessel based on additional information gathered throughout the interview (e.g. number of hauls per day, total days fished per trip, number of trips per month, and total months fished per year). We attempted to minimize the biases in values fishers’ provide about frequency, duration and seasonality of fishing activities (e.g., number of days fished per month) by asking questions during the interview about number and frequency of vacation, rest and sick days as well as seasonality of events (Vaske et al. 2003, Schmidt et al. 2015).  While some fisheries studies have suffered from non-response bias (Connelly and Brown 1995, Rao et al. 2004, Moore et al. 2010), we used short (15 min) semi-structured interviews, where respondents were free to skip questions or stop the interview at any time, a technique recommended to reduce non-response bias (Tarrant et al. 1993, Schmidt et al. 2015). Additionally we employed clearly defined time periods by asking follow up questions about fisher reported time periods and asked about typical catch to aid respondents to recall fishing   8 effort and avoid telescoping and reporting “best” catch scenarios (Coughlin 1990, Morwitz 1997).   2.3.2 Calculating annual catch rates per fisher First we calculated annual catch rates for each fisher by scaling up his/her reported catch rate with his/her reported frequency of fishing effort (e.g. daily catch rate × total days fished per month × total months fished per year). We executed this analysis for all reported catch rates including multiple catch rates reported for each fisher. For example if a fisher reported a catch rate in both per day and per month, we calculated two annual catch rates for that fisher.    2.3.3 Differences across reporting time periods: among different fishers, each reporting only one time period (Among Fishers) We compared annual catch estimates among fishers who each chose one time period in reporting their catch for each gear type (n = 54 trawlers, n = 51 shrimp gillnet fishers, n = 70 fish gillnet fishers, n = 80 crab gillnet fishers, and n = 18 other gillnet fishers). Our selection enabled us to maintain the independence of each data point and also avoid subjectively choosing any one reporting time period for fishers that reported multiple metrics (e.g. per haul, per trip, per month). Additionally for gillnet fishers, we pooled those fishers that responded in per haul, per-day and per-trip, since more than 90% of the fishers reported setting and retrieving their net once per day, with fishing trip duration of one day.   For each gear type, we then compared the differences in annual catch estimates among the different time periods. Since the shape of the distribution for each time period was not identical   9 (based on Kolmogorov-Smirnov test), we could not statistically compare the mean or median among groups (McDonald 2014). Instead, we used a Kruskal-Wallis test to compare the mean ranks among groups (McDonald 2014). Our null hypothesis was that the mean ranks of the annual catch estimates based on different time periods were the same. Then we used the Dunn’s test for post-hoc multiple pair comparisons (Zar 2010) to identify where any such differences occurred.    We also developed generalized linear mixed models to examine the effects of reporting time period on the annual catch, with considerations of two confounding factors: gear type and fishers’ location (province) as random variables. The link function of the model is a log function, since the distribution of the response variable (annual catch estimate) is right-skewed. We then examine the significance of the fixed effect (time period) by comparing our model to a null model (without time period as the fixed effect) using the likelihood ratio test.  2.3.4 Differences across reporting time periods: across the estimates from an individual fisher who reported more than one time period (Within Fishers) To examine the possibility of recall biases, we compared a fisher’s reported catch across all the time slices s/he gave. First, we selected the fishers who gave multiple catch rates based on different time periods for each gear type (n=14 for gillnet and n=28 for trawler). We then calculated the ratio of scaled annual catch rates for every pair of reported time periods. For example, if a fisher reported his/her catch in per-haul, per-day and per-trip, we calculated the ratio of annual catch from (i) per-haul versus per-day, (ii) per-haul versus per-trip, and (iii) per-day versus per-trip. Our hypothesis was that if fishers were not prone to recall bias, their catch   10 estimates scaled from each metric would be consistent, and thus the ratios will be close to one. Conversely if fisher’s were prone to recall bias, the ratios comparing catch estimates would differ significantly from one, and we used a bootstrap test with 1000 iterations to test for this significance (Efron 1979). In this analysis, we pooled all samples from different types of gillnet fishers to have a sample size large enough for hypothesis testing.  2.3.5 Comparing catch rates with external datasets  We compared our interview catch rates with external datasets to see which time period(s) best capture the annual catch of Thai seahorses. In this analysis, we used all the data, pooling data from the Among Fishers and the Within Fishers information. We aimed to examine two hypotheses: (1) there is a time period that is “most appropriate” for fishers to use in reporting seahorse catch rate, as was the case in a study on wildlife hunting (Golden et al. 2013); (2) the mean and median should be calculated across all reported catch rates, regardless of time period, to account for the variations among and within fishers.    We compared our catch rates by (1) time period and by (2) mean and median of all reports to the catch rates obtained by two external datasets for Thailand’s seahorse catch: landings catch rates (Laksanawimol et al. 2013) and government research trawls (Thai Department of Fisheries, unpublished data). First, we accessed annual catch estimates from landings in Laksanawimol et al. (2013), and compared those to the estimates in our study. The seahorse landings assessed in Laksanawimol et al. (2013) were also bycatch from non-selective fishing gears (trawl, gillnet, and pushnet), with the data collected from 2011-2012. To account for spatial variation in reported catch rates, we only compared catch rates from Trat province, where both datasets   11 overlapped. Since Laksanawimol et al. (2013) reported in kilograms, we applied the conversion rate of 3.13 g per seahorse (Perry et al. 2010) to convert the estimates into number of seahorses. Only mean values, but not medians, were provided in this landing data (Laksanawimol et al. 2013), so we could not compare median values. We compared our catch rates reported by trawlers to the annual landings from middle-scale fisheries (shrimp or squid trawlers) reported in Laksanawimol et al. (2013), and compared our reports from gillnet fishers (pooled all types together) to the landings from artisanal fisheries (gillnet and pushnet) reported in Laksanawimol et al. (2013). Second, we corroborated our fisher-reported annual catch rates with catch rates from a three-year dataset of government research trawls from the Thailand Department of Fisheries (DoF) (see Corroboration with governmental trawl survey data section in supplementary material). Since the DoF trawl surveys were conducted by otter trawl, we only compared this data to our interviews from otter trawl fishers (n=34). Trawl surveys were executed by DoF in 2010, 2012 and 2013 at pre-determined sampling locations throughout the Andaman Sea and Gulf of Thailand (Fig. A1). These locations were originally intended for the purpose of sampling commercially exploited fish species, but recording the presence or absence of seahorses for each research trawl began in 2010 (Table A1). Each location was sampled four times per year using a 23.5 m otter-board trawl research vessel, with trawl speed set at 2.5 nautical miles/hour. Research trawls took place within each location for one hour. In order to compare this effort to our reported trawl fisher effort (mean = 4 hours/haul), we multiplied the DoF catch rate by 4 (Table A2).     12 2.3.6 Implications for other seahorse studies We conducted a review of published literatures to investigate the implications of reporting various time periods in the seahorse catch estimates of previously published studies. We first collected all available information on seahorse catch rates derived from field-based fisher interviews, and found that the time periods fishers used to report their catch rates varied in the literatures. We then summarized the data and identified the most highly reported time period across the literature. Next, we wanted to determine if any previously published studies had chosen a specific time period to scale up annual catch rates when more than one was available, and how this choice would affect catch estimates. To do so, we identified the number of studies that reported catch rates by multiple time periods, where one of the reporting time periods was by haul (n=3). Since “per haul” is the smallest time period ever used in seahorse catch surveys, we expected that if recall bias exists in previous research, the estimates based on “per haul” would cause the largest effects. From these results, we selected the papers where the fishing effort metrics (e.g. days fished/month, months fished/year) used to scale up the catch rate to annual catches were reported (n=2). We then re-calculated the annual catch of seahorses based on the per haul time period for these papers. For these calculations we used the same number of reported total boats, and days or months fished as used for the original annual catches to determine our new annual catches based on haul. We then compared the scaled up annual catch to the original estimate reported in the literature.    13 2.4 Results 2.4.1 Differences across reporting time periods: among fishers reporting only one time period (Among Fishers) For both commercial trawlers and small-scale gillnets fishers, the estimated annual seahorse catch from shorter time periods was higher than when fishers reported using longer time periods (Table 2.1-2.3 & Fig. 2.2). The median of annual catch of fishers reported per-haul was 5,000 (trawlers) to 260 (gillnet fishers) times higher than the ones reported per-year (Fig. 2). For trawlers and most gillnet fishers, the Kruskal-Wallis test and post-hoc Dunn’s test agreed that the annual catch rates differed significantly according to which time periods the respondent used, with lower catches for longer time periods (Table 2.2 & Table 2.3). Although such difference was not statistically significant for the other gillnet fishers, those with no specific target species mentioned, the pattern among time periods still existed (Table 2.3 & Fig. 2.2). The variances of the annual catches per vessel were higher for the reports in shorter time periods (Table 2.1 & Table 2.3).   The pattern of inferring decreasing annual catch rates with increasing time period was also found when we considered both gear type and fishers’ location as confounding factors using the generalized linear mixed model. When comparing our model to a null model (without time period as the fixed effect) using the likelihood ratio test, we found that the reporting time period had a significant effect on the annual catch estimate (χ2(4)=909869, p<0.0001).    14 2.4.2 Differences across reporting time periods: within individual fisher who reported more than one time period  (Within Fishers) We found that even for the same fisher, time periods of per haul or per day led to higher annual catch rates than estimates from per trip or per month (Table 2.4 & Fig. 2.3). This result was found in both commercial trawl fishers and small-scale gillnet fishers. However, for trawl fishers, catch rates based on per haul were not significantly different to the ones based on per day (p=0.19), and per-trip rates were similar to per-month (p=0.27) (Table 2.4).  2.4.3 Comparing catch rates with external datasets  No single time period was corroborated as the ‘most suitable’, by comparing our data to previously published landings data or government trawl survey data (Table A3, Table A4). For trawl fishers, the corroboration with landings data could only be conducted for the Trat province (Laksanawimol et al. 2013). Our only per trip estimate (246 seahorses/vessel) in Trat came from just one fisher, though it was similar to the landings data (219 seahorses/vessel) (Laksanawimol et al. 2013). We did not find any time period matched well with government trawl data (Table A3, Table A4). The mean annual catch rate estimated from the government trawl surveys, either considered the survey efforts of two coasts equally (5,079 seahorses/year) or unequally (2,771 seahorses/year) (See Supplemental Material), was three times higher than the reports based on per-day but was only 40% of the per-haul reports, and double the per-trip reports and 20% of the per-haul reports, respectively (Table 1). The median annual catch rate of the governmental trawl surveys was zero seahorse/year, which was closest to the interview reports based on per-year (median=2 seahorse/year), though we only had two otter trawl fishers reported in per-year (Table   15 1). For small-scale gillnet fishers, no single time period was similar to the landings data (Table A4).  The mean across all trawl fishers’ reports (4,977 seahorses/year) was similar to the mean catch rates from governmental trawl surveys, if both coasts were weighted equally (5,079 seahorses/year) (Table A3). Comparisons using median calculations did not yield similar results for annual catch rates from the DoF government trawl surveys and our trawl fisher reported catch rates (Table A3).   2.4.4 Implications for other seahorse studies We found that 65% of the catch rates reported in previous seahorse studies using fishers’ interviews reported in per-month or per-year (Table A5). For those studies reporting multiple catch rates including per haul (n=2), annual catch estimates with per haul values were higher than the estimates based on other reporting time periods (Table A6).  2.5 Discussions We found that fishers that reported for a shorter time period tended also to report a higher annual catches, and our Within Fisher analysis further suggests a potential for recall bias across reporting time periods. Our findings complement conclusions from recreational fishing and hunting studies, in which respondents tended to mis-report their fishing/hunting estimates when recalling over a unsuitable time period (Fisher et al. 1991, Connelly and Brown 1995, Golden et al. 2013). Unlike previous wildlife studies, we could not conclude that any single time period as the “most suitable” for reporting seahorse catch rates based on comparison to external datasets   16 (Golden et al. 2013). However, our corroboration of commercial trawl survey data suggests that summarizing data across all reports, with mean values, may be applied with caution.   There are some potential mechanisms that may explain the large variations in fisher reported catch rates based on different reporting time periods (Coughlin 1990, Denzin and Lincoln 2011). First, for the differences in catch rates among fishers (Among Fishers) may represent variations in catchability of seahorses, it could be due to natural variability of fishers in time and space and the patchy distribution of seahorses (McShane 1996, Vincent et al. 2011, Barbeaux et al. 2014). In other words, it is possible that fishers who catch fewer seahorses (a more rare event to them) tended to use a longer time period in reporting. However, this cannot explain the discrepancies in multiple catch rates reported by the same fisher, as found in our results (Within Fishers). A fisher who reported a larger annual catch when using a larger time period can be resulted from two reasons. The first explanation may be that fishers can better recall a more recent event than one that happened farther in the past (Tarrant et al. 1993,). If this explanation is true, then catch rates from a longer time period would underestimate current catch rates, which may not capture the impact of fisheries on data-poor species (Johannes et al. 2000). The second mechanism may be that fishers have a tendency to under-report zero catch, which would cause catch rates from shorter time periods to be over estimates, as found in other studies (Cannell et al. 1977, Vaske et al. 2003, Golden et al. 2013). Considering that catching seahorses may be an irregular event, especially in small-scale fisheries (O’Donnell et al. 2010), this bias may exist in our small-scale fisheries results, which highlights the importance of corroborating data (O’Donnell et al. 2012, Golden et al. 2013). Our corroboration of results proved inconclusive for small-scale fisheries, suggesting that there was no one dominant mechanisms driving our results.     17  Our corroborations with external datasets suggested that analyzing our fisher reported data including all reports may be a better reflection of true catch rates than any one reporting time period, which contrasts previous recreational fishing and hunting research (Golden et al. 2013, Zarauz et al. 2015). Such results support other research recommending that all interview data should be included to take into account the variations (Grant and Berkes 2007, Barbeaux et al. 2014). Despite this, we should bear in mind the caveats of comparing data collected by different methods. For example, fishery-dependent data and scientific surveys may differ at collection times and locations (Hilborn 1985, Babcock and Pikitch 2011), complicating corroboration attempts with external datasets. However, in our study, seahorses are a non-target catch and the decisions about where and when to fish are not made in relation to seahorses. This enhances the reliability of comparing our fishery-dependent data to the trawl surveys, as the governmental trawl surveys also focus on common target species in Thailand rather than seahorses.  We believe current estimates of seahorse bycatch (Choo and Liew 2005, Salin 2005, Baum and Vincent 2005a, Giles et al. 2006, Perry et al. 2010, Lawson et al. 2016) are likely under-estimated for three main reasons. First, we found that the majority of previous seahorse catch estimates were based on longer reporting time periods (per month or per year), and with the potential for recall bias would result in a smaller catch estimates (Baum and Vincent 2005a, Lawson et al. 2016). Second, global catch estimates of seahorses have typically not included estimates of illegal, unreported, and unregulated (IUU) fishing, which are often the largest uncertainty in catch estimates (Agnew et al. 2009, Jacquet et al. 2010, Varkey et al. 2010). For example, in Thailand, IUU fishing is a significant challenge, and historically there have been   18 high levels of uncertainty in the number of commercial and small-scale fishing fleets operating in Thai waters (Teh et al. 2015). Lawson et al. (2017) did not attempt to incorporate IUU fishing estimates into their global analysis, suggesting that the 37 million estimate is likely an underestimate. Finally, underreporting is commonly seen in self-reporting of catches and in fisher interviews, since respondents may downplay catches for fear of additional restrictions (Filion 1981, Jones et al. 2008, Rist et al. 2010). Many seahorse studies (Baum and Vincent 2005, Perry et al. 2010, Salin 2005, Giles et al. 2006) rely on fisher interview data because seahorse catch is typically not recorded in log-books, government officials or in fishery independent surveys, thus leading to the potential for under-reporting.  In order to reduce the recall bias described in this study, we would improve the data collection process with the following adjustments. The first is to carefully design interview questions to obtain information about zero-events (Vaske et al. 2003, Golden et al. 2013), which provides a better understanding of reported catch rates. Second, conducting a short, fishery-dependent pilot study, such as port sampling, logbooks or onboard observations, could complement fisher interviews (ICES 2010). This type of pilot study would allow for cross-validation of fisher-reported catch rates to highlight any recall bias due to reporting time period. Both approaches would assist in reducing bias and increasing reliability of local knowledge for more robust management (ICES 2010).   Despite data uncertainty and the potential for bias, as in our case, it is still necessary to make management decisions for species conservation and fisheries management (Johannes et al. 2000, Regan and Ben-Haim 2005). Although fisher interview data is far from perfect, it could serve   19 well as a baseline for management especially in data-poor situations (Thornton and Scheer 2012), and furthermore promotes inclusiveness in the management process (Berkes et al. 2000). Additionally, information from fishers’ knowledge is typically cost-efficient to acquire, especially in developing countries (Ban et al. 2009, Moore et al. 2010, Thornton and Scheer 2012). As we can never wait for perfect and complete data to take action, information from fishers’ knowledge is the most cost-effective starting point for adaptive management (Walters and Holling 1990, Berkes et al. 2000).     1 Table 2.1 The median and mean of annual catch per vessel scaled up from various reported catch rates for trawlers reporting one time period. Kruskal-Wallis tests showed the mean ranks of scaled catch estimates were significantly different (p<0.01) among different reported time periods.  Trawlers’ scaled mean annual catch rates  Median  Mean n SD Haul 5,200  33,468 11 56,958 Day 1,680 1,782 9 2,212 Trip 246 735 26 952 Month 105 158 4 135 Year 1 3 4 5 Overall 480 6,785 54 26,783 χ2 (df=4) 32.51 P-value <0.01**   2 Table 2.2 Results of pairwise comparisons (Dunn test) for annual catch of trawlers reported by time period. Numbers show the z-test statistics with p-values in the brackets for each pair of time periods. Significant values (<0.05) are italicized.   Haul Day Trip Month Day -1.51 (0.07)     Trip 3.69 (<0.01)  1.99 (0.02)   Month 3.92 (<0.01)  2.48 (<0.01) -1.34 (0.08)  Year 4.73 (<0.01)  3.50 (<0.01) 2.49 (<0.01) 0.87 (0.18)   3 Table 2.3 The median and mean of annual catch per vessel scaled up from various reported catch rates for gillnet fishers reporting one time period. Kruskal-Wallis tests showed the mean ranks of scaled catch estimates were significantly different (p<0.01) among different reported time periods. The results of the pairwise comparisons (Dunn test) show the z-test statistics with p-value in brackets.  Shrimp gillnet Fish gillnet Crab gillnet Other gillnet  Med.  Mean n SD Med.  Mean n SD Med.  Mean n SD Med. Mean n SD Day 261 425 10 505 564 1,473 18 2,574 414 649 17 750 42 42 2 60 Month 18 45 28 90 30 50 13 65 18 44 34 44 3 18 5 31 Year 1 3 13 4 0 0 39 2 0 2 29 2 0 2 11 4 Overall 18 109 51 275 2 389 70 1,430 12 151 80 420 0 11 18 25 Kruskal-Wallis Test results χ2 (df=2) 28.02 57.29 52.78 2.26 P-value <0.01** <0.01** <0.01** 0.32 Dunn Test results  Day×Month 2.60 (<0.01) 2.26 (0.01) 2.54 (<0.01) 0.15 (0.44) Day×Year 5.20 (<0.01) 7.31 (<0.01) 6.78 (<0.01) 1.06 (0.14) Month×Year 3.66 (<0.01) 3.94 (<0.01) 5.31 (<0.01) 1.27 (0.10)   4 Table 2.4 Mean ratio of each pair of scaled annual catch by reporting time period for fishers who reported multiple time periods. The mean of all ratios are shown, with sample sizes (n) of how many fishers reported on that pair of time periods. P-values <0.05 (bootstrap test) are italicized, representing mean ratios significant different to 1.  Trawlers Gillnet  Haul Day Trip Day  Ratio n p Ratio n p Ratio n p Ratio n p Day 1.33 6 0.19          Trip 4.80 11 <0.01 2.36 5 <0.01       Month 10.8 2 -- 2.50 3 0.04 1.07 3 0.27 8.30 11 <0.01 Year  15.03 3 <0.01   5  Figure 2.1 Commercial and small-scale fisher interview locations in Thailand.  6  Figure 2.2 Comparison of scaled annual catch reported by time period among fishers reporting only one time period. Seahorse catches from (a) trawlers, (b) gillnets for shrimp, (c) gillnets for fish, (d) gillnets for crab, and (2) other gillnets are shown, with number of fishers (n) reported their catch on each time period.       7  Figure 2.3 Pairwise-comparison of scaled annual catch reported by time period from (a) trawlers and (b) gillnet fishers reporting more than one time period. For each fisher, we calculated the ratio of the two scaled annual catches for each pair of the time periods. If the catch estimates from two time periods are similar, the ratio should be close to one (dash line). Abbreviations for each time slice are: H: haul, D: day, T: trip, M: month, and Y: year. H:D represents the ratio of annual catch estimate from per haul to estimate from per day.       8 Chapter 3: Implementation of CITES listing stimulates changes in trade: the case of dried seahorse trade in Thailand  3.1 Synopsis Exploitation for trade is one of the biggest threats to many species, especially for marine fishes. Trade regulations should, therefore, be effective in helping conserve marine fish populations. The Convention of International Trade in Endangered Species of Wild Fauna and Flora (CITES), one of the few multilateral environmental agreements with enforcement capacity, has embraced a number of marine fishes in recent years. However, the impacts of such measures on wildlife trade have rarely been assessed. We conducted a case study of the dried seahorse (Hippocampus spp.) trade in Thailand to understand the trade of these species under CITES regulations. We carried out 203 semi-structured interviews with traders to estimate the economic scale of Thai seahorse trade, and compare perceived changes with official trade datasets. Even though most seahorses were incidentally caught, we estimated that dried seahorses could be worth US$26.5 million per year for fishers. However, the total declared annual export value was only around US$5.5 million, and had decreased to US$1 million in 2013. Considering the economic value of seahorses, the large discrepancy between declared export volumes and catch estimates suggested that trade may be underreported. While official data shows the export volume decreased after the implementation of CITES listing in 2005, our respondents did not report a similar trend. In contrast, the prices of seahorses were reported to be increasing. Our study highlights the economic importance of marine fishes captured as bycatch and the importance of international and domestic management measures for the trade of bycatch species.   9  3.2 Introduction Many wildlife populations are in crisis because of human activities, especially for marine species. Biodiversity is being lost at an unprecedented rate, locally and globally, because of anthropogenic effects such as over-exploitation, habitat destructions and climate changes (Aarts and Poos 2009, Abensperg-Traun 2009, Costello et al. 2012). A dominant concern is that exploitation for trade affects millions of animals and plants every year (Smith et al. 2009c). International declared wildlife trade was estimated to be worth about US$332.5 billion in 2005 (Engler 2008), and the high monetary value provides incentives for people to keep extracting animals and plants. Consequently, the extinction risks for traded wildlife species have been accelerating (Lenzen et al. 2012b). Among the hundreds of millions of plants and animals that are traded every year, fisheries products comprise one-quarter of the total in terms of value (Engler 2008). Annually, about 79.7 million tonnes of marine species are extracted from the oceans, and traded for food both domestically and internationally (FAO 2014), contributing more than US$80 billion to the world economy (Willman et al. 2012). Approximately 28-33% of all fish stocks are overexploited because of unsustainable fishing, and declines in abundance continue in many other stocks (Branch et al. 2011). Finding ways to ensure the sustainable use of wildlife, including marine fishes, has become an urgent conservation challenge.  Trade measures have commonly been used to improve the sustainability of wildlife trade, and holds enormous potential to supplement management measures for marine fishes. Trade regulations - such as tariffs, quotas and bans - aim to control the supply of and/or demand for endangered species (Challender et al. 2015a). Even trade treaties that advocate trade   10 liberalization, such as the General Agreement on Tariffs and Trade (GATT), state that Signatories shall be allowed to adopt measures that are ‘necessary to protect human, animal, or plant life or health’ (Wold 1996). Recent multilateral free trade agreements (e.g., Trans-Pacific Partnership) even require Parties to implement specific multilateral environmental agreements, such as the Convention of International Trade in Endangered Species of Wild Fauna and Flora (CITES) (Lurié and Kalinina 2015). While international trade regulations have been widely used to protect terrestrial species (e.g., EU’s import bans for birds and tariffs for timbers in many countries), conservation measures undertaken for marine fishes have mainly focused on restricting fisheries, with relatively few cases addressing trade (although note examples such as the US import restrictions on shrimps caught by nets without turtle excluder devices) (Vincent et al. 2013).   Recently, a number of economically-valuable marine fishes have been embraced by CITES, the biggest multilateral agreement for international trade with a conservation purpose (Vincent et al. 2013). The species covered by CITES are listed in Appendices to the Convention, according to the degree of protection they require. All species in Appendix I are threatened with extinction because of international trade, and the trade of these species is basically prohibited. Species on Appendix II are or may become threatened by international trade. They can only be exported with permits that national government grants after ensuring the specimens are legally sourced and trade will not endanger wild populations – termed making a Non-Detriment Finding (NDF) (CITES Resolution Conference 16.7 2013). If a Party fails to meet these obligations, the export of this species may be suspended (e.g., exports of Hippocampus kuda from Vietnam, CITES, 2016). Thus, CITES is the only multilateral environmental agreement with enforcement capacity   11 (Ardron et al. 2014). The outcomes of CITES implementation for the conservation of marine fishes desperately needs to be examined.  CITES implementation can have positive effects by generating monitoring and management for listed species at the national level (OECD 2000). To make the positive NDFs required to permit trade, Parties need to design and implement domestic management measures that can help secure healthy populations (Dickson 2013). Common national implementations of CITES listing including domestic bans on extraction (Hutton 2013), trade quotas (Raymakers and Hoover 2002, Sadovy et al. 2007), and size restrictions (Mejía et al., 2008). Improvement of ranching and captive breeding techniques have also been stimulated by CITES as these can allow traders to obtain permits more easily (Robinson et al. 2015). For marine species, CITES listings have also stimulated reform in fisheries and trade management. For example, the stock assessment of humphead wrasse (Cheilinus undulatus) led to the establishment of a trade quota in Indonesia (Sadovy et al. 2007, Sadovy 2010). In another example, Jamaica introduced its management plan for queen conch (Strombus gigas), including restrictions on the number of vessels and closed season, in order to respond to CITES’ requests (Aiken et al. 1999, Catarci 2004).   National management strategies in response to CITES measures are usually cognizant of the need for biological sustainability, but rarely consider the socio-economic implications of such strategies, which in turn could change expected outcomes (Challender et al., 2015a; Velázquez Gomar and Stringer, 2011, Smith et al 2011). Management measures that reduce stakeholders’ (e.g. fishers’ and traders’) revenue may lower their willingness to participate in conservation actions and discourage them from following regulations (Roe et al. 2002, Sorice et al. 2013). In   12 addition, while CITES listing has reduced consumer demand for some species by generating conservation awareness (Challender et al., 2015a), listing has also provoked undesirable consequences such as growing demand in other endangered, high-value species (Challender et al. 2015a). Diverse responses of the traders and markets to trade interventions are associated with the cultural values of the species, economic status of the stakeholders, and many other social-economic factors (Roe et al. 2002, Challender et al. 2015a). At CITES Conference of the Parties 13 (2003), Parties agreed that the social-economic perspectives of the exporting country (e.g., enforcement and rural population’s livelihood) should be considered while implementing CITES (Velázquez Gomar and Stringer 2011). Understanding the types of use, trading systems, and the scale of the industry is a critical step towards achieving effective natural resource management (Smith et al. 2011).   As some of the first marine fishes listed on CITES since its inception, the 41 seahorses species are setting precedents for other marine fishes protected by CITES, yet the effect of CITES on their trade is still unclear. The whole genus of seahorses was listed in CITES Appendix II in 2002 (implemented in 2004) because of large-scale and intensive trade in wild populations. CITES data reported that around 3.3-7.6 million seahorses were traded globally from 2004-2011 (Foster et al., 2016). However, data generated by detailed trade surveys estimated that the global catch of seahorses were several times higher than the CITES records, at around 37 million individuals (Lawson et al. in press). The majority of seahorses are traded dry (Foster et al 2016) for use in traditional medicines and curios (Foster and Vincent 2005). Most exported seahorses are caught in non-selective fishing gear; in many places the biggest threat for seahorses is their incidental capture during trawling (Lawson et al., in press).  Critical to evaluating the impact of   13 CITES on seahorses is the existence of Customs data for seahorse imports, which started before the CITES listing and provide an invaluable opportunity for us to identify the changes in trade because of the CITES intervention.   Thailand has long been the biggest exporter of wild, dried seahorses, exporting approximately 88% of the total numbers reported in the CITES database (Foster et al., 2016). When CITES scrutinized Thailand’s implementation of the Convention, it deemed Thailand’s exports of three seahorse species (Hippocampus kuda, H. kelloggi, and H. spinosissimus) to be of Urgent Concern. In 2012, CITES issued Thailand with formal recommendations to address its challenges in making positive NDFs for those seahorse species. CITES Appendix II species are commonly regulated through nationally specific export quotas (Challender et al. 2015a), and so as a precautionary measure, Thailand responded to the CITES recommendations by suggesting a maximum export volume (1,500 kg) to exporters. Such a quota, though not mandatory, sought to ensure export levels did not increase while additional support for an NDF was undertaken. This quota was only 10% of pervious export volume, and was considered as the maximum sustainable yield of seahorses (Phoonsawat et al. 2015). In 2016, Thailand suspended their seahorse exports. However, no study has been done to understand the restrictions on exports to the fisheries and trade. The last national seahorse trade survey in Thailand was conducted in 1998-1999 (before the CITES listing: Perry et al., 2010), so there is a need for post-CITES data to examine the effect of trade regulations.   The objectives of this study were to generate and share new knowledge about Thailand’s trade in dried seahorses since the CITES listing, and explore how the national trade of these valuable   14 marine fishes has changed with trade regulations. To meet these objectives, we gathered data in 2013-14 after seahorses had been listed in CITES Appendix II. The findings were compared to official statistics for exports, which included pre-CITES period, to assess the impact of CITES implementation. We are interested in two questions: (1) what is the economic scale of the seahorse trade in Thailand; and (2) how might trade have changed with CITES interventions? The three CITES interventions we consider are: CITES listing in 2002, CITES listing implementation in 2004, and the voluntary export volume suggested by the Thai government in 2012 (the export suspension, which starts in 2016, was out of our data scope). We hypothesized that the trade volume would decrease only after the domestic policy was applied, while the price would increase after the CITES listing. The findings of this study will assist the Thai government in developing a management strategy for seahorse trade, while shedding light on the social-economic effects of trade regulations on wildlife.   3.3 Methods To meet the first objective of this research, which was to generate new information of current trade status of dried seahorse trade in Thailand, we conducted a total of 203 semi-structured interviews with people whose work was related to seahorse trade (e.g., fishers, traders, government officers) from 2013 to 2014.  For the second objective, to determine whether the trade changed with the three stages of CITES implementation in Thailand, we examined trade volume and prices reported in interviews, and compared our interview results to three external datasets: export and import records from the CITES trade database for 2005-2012 (UNEP-WCMC 2013a), and Customs statistics from two of the largest seahorse importers, Hong Kong SAR (1998-2012) and Taiwan (1983-2012).    15  3.3.1 Trade surveys - data collection Our surveys covered 13 Thai provinces (Fig. 3.1) where trade was known or likely to occur based on two sources: (1) earlier studies of the seahorse trade in Thailand (Perry et al. 2010, Laksanawimol et al. 2013, Phoonsawat et al. 2015) and (2) suggestions from respondents during the surveys. In order to find respondents who were involved in the dried seahorse trade, we undertook snowball sampling in which one respondent indicated other potential respondents (Biernacki and Waldorf 1981). The first and second authors (Taiwanese and Thai citizens respectively) conducted the field surveys. During each interview, questions and answers were translated between Thai and English (n=172, by TCK and PL), conducted in Thai (n=24, by PL) or conducted in Mandarin or Hokkien (n=7, by TCK). All research was approved by the University of British Columbia’s Animal Care Committee (5706-12) and Human Behavioural Research Ethics Board (H12-02731).   During the interviews, we asked questions that included but were not restricted to past and present use of seahorses, trade routes, trade volumes, prices, and seahorse characteristics (colour, spiny/non-spiny, size, etc.). Data were recorded in the units given by respondents but standardized during the analysis for ease of comparison (see below). In order to understand the perceived changes in seahorse trade, traders were asked to either recall the seahorse trade volume or/and price of their first year working in this industry, or of a certain year that was meaningful to them (e.g., the year before export permits were required or the year they started trading seahorses). We used triangulation to cross-validate the information received within interviews by asking the same questions in different ways and comparing the answers within and among   16 interviews, and across trade levels. Each interview took from 5 minutes to 1.5 hours, depending on the engagement of the respondent. We also weighed and measured the height (coronet to tail) of 126 seahorse specimens that were found in trade.  3.3.2 Trade surveys - data analysis After the interviews, we categorized the respondents according to their role in, or involvement with, the seahorse trade, including fishers, buyers, consolidators (wholesalers), domestic retailers, exporters, government officials, and other experts (Table 3.1). Fishers and traders were further categorized into different trade levels (Fig. 3.2). For example, fishers who caught seahorses directly were defined as “level 1”, collectors that bought directly from fishers were “level 2”, and so on. Note that one respondent could occupy more than one level in the trade chain, and if one did, we defined his/her trade level by the highest position he/she occupied in the trade chain.   We collected information on annual seahorse trade volume and price for each trade level. Reports from individual respondents on trade volumes for discrete time periods (i.e., per week, trip, month, or year) were scaled to annual estimates based on the frequency of trade activity (e.g., number of purchases in a month). Since the price of seahorses varied with size, we standardized the reported prices to Thai Baht (TBH) per gram dried weight.  To compare the price across sizes, we divided seahorses into three size-classes: 300-800 seahorses per kilogram (size class Small), 101-300 seahorses per kilogram (size class Medium), and 40-100 seahorses per kilogram (size class Large).  To compare the prices reported in a different currency (e.g., exporters might report the price in USD), we used the mean exchange rate of TBH and USD   17 from January 2012 to March 2014 (0.032 USD = 1 TBH) (http://www.usforex.com, accessed April 29th 2014). We carried out all analyses based on seahorse dry weights, converting numbers of individuals to weight using 1 dried seahorse = 3.16g; this was the mean weight of individual seahorses collected during the surveys (n=126, SE=0.16).   3.3.2.1 Estimate economic scale To quantify the economic value and effects of CITES on the trade, we used four metrics to determine the economic scale of the dried seahorse industry in Thailand: (1) value generated when the fishes were extracted from the water, termed direct expenditures or gross output; (2) gross value of annual domestic trade; (3) gross value of international trade (i.e. annual exports); and (4) potential number of traders involved in every level of the supply chain. We did not quantify the profits of the seahorse trade because we lacked data about costs. Therefore, we quantified the gross output and export values instead.   Direct expenditure (1) was estimated as the total catch of seahorses in a year (91 tonnes after conversion from number of individuals; Aylesworth et al. in review) multiplied by the mean ex-vessel price reported by fishers (level 1) and primary buyers (level 2). The gross values of annual domestic trade (2) were calculated by multiplying the mean selling price, the mean trade volume of individual domestic TCM retailers, the total number of registered TCM retailers.  To estimate domestic trade volumes, we standardized the reported sold/purchased volume of seahorses into annual quantities for each respondent. The total number of TCM shops in Bangkok (n=491) was accessed from the Bangkok Medicine Trader Association (http://www.thaicn.net/zyzy.html, accessed April 10th 2013). Then, we multiplied the mean price per gram by the total trade   18 volume for retailers to calculate the gross value of domestic trade and exports. The gross values of annual export trade (3) was calculated as the mean selling prices multiplied by the export volume reported in the CITES trade database.  To understand how many people were involved in the dried seahorse trade (4), we estimated the potential number of traders in each trade level, except for level 1, for which we calculated as the number of trawl vessels instead of number of fishers. The calculation process is detailed in the Appendix B. We created two estimates of the number of people involved in the trade: (i) starting at the top of the trade chain (total exports) and working down to number of trawlers, and (ii) starting with an annual catch estimate (Aylesworth et al. in review) and working up to number of exporters. Since respondents might report purchase volume, sell volume, or both, we applied two methods to calculate the number of traders in each level (using trawlers as level 1), except for level 3 where we lacked data. In Method 1, we considered both trader-reported annual purchasing volumes and selling volumes. In Method 2, we considered either selling volume or purchasing volume, or the larger of the two reported volumes if a trader reported both. These two methods were applied for both top-down and bottom-up calculations. For the top-down calculations, we started with the export volumes reported to CITES in 2011 and 2013; they straddled the arrival of the voluntary export quota in mid 2012. For bottom-up calculation, since the catch estimate was based on known number of fishing vessels (Aylesworth et al. in review), we skipped level 1 (trawlers) and started our estimate from level 2 (primary buyers).     19 3.3.2.2 Estimate income from dried seahorses We estimated the mean income from dried seahorses for individual fishing vessels and exporters (the bottom and the top of the supply chain). For a fishing vessel, the income from seahorses was calculated by dividing the gross output (value of total catch) by the total number of registered boats (considered all gears that catch seahorses, Aylesworth et al. in review); this calculation assumed all catch goes into trade. We also estimated mean income for vessels based on total declared export and domestic trade volumes, reflecting the possibility that not all caught seahorses enter trade. Mean income per vessel was calculated for commercial and small-scale fisheries separately, by considering the proportion of catch from each type (68% from commercial and 32% from small-scale; Aylesworth et al. in review). The mean income per exporter was calculated as mean reported export volume multiplied by mean reported export price.  3.3.2.3 Perceived changes in trade volume and prices over time and with CITES Implementation We collected the historical trade volumes and selling prices to explore changes in seahorse trade over time for each trade level. To account for inflation rate, we corrected historical prices by scaling them with the relative consumer price index (CPI) in Thailand for 2014 (CPI of 2014=104, Bank of Thailand 2016).    3.3.2.4 Examine changes in international trade using external datasets In addition to interview data, we also used three sets of official trade records to examine changes in the international trade of Thai seahorses. The three data sets were: 1) CITES trade database   20 (https://trade.cites.org, UNEP-WCMC, 2013), 2) statistics from the Census and Statistics Department of Hong Kong SAR and 3) Customs Administration records from Taiwan (https://portal.sw.nat.gov.tw/APGA/GA03, accessed at December 7th, 2015). CITES trade data were downloaded on September 9th 2015 for all dried trade records involving Thailand from 2004 to 2013. Since the CITES listing for seahorses was implemented in May 2004, data for 2004 may reflect only partial trade volumes for that year (Foster et al 2016). While the CITES trade database contained both import and export data, we used the “exports reduced from imports” as defined in Foster et al 2016. Hong Kong SAR’s Census and Statistics Department recorded dried seahorse imports (in kg) and prices into Hong Kong from 1998-2014, and Taiwan’s Customs Administration recorded dried seahorse imports (in kg) to Taiwan from 1983-2014, and prices from 2002-2014 (although the prices of imports from Thailand were missing in 2013 and 2014).   3.4 Results  3.4.1 Trade structure We found at least five trade levels for the dried seahorse trade in Thailand, with any individual potentially operating in several levels. Fishers (level 1) (n=98) reported selling dried seahorses to primary buyers (level 2) at ports (Fig. 3.2). Most level 2 traders we interviewed lived near the ports so they could easily visit whenever trawlers landed their catches (n=16 of 22); some even maintained grocery shops beside the port as a trading venue for fishers (n=7 of 22). Some level 2 traders bought seahorses along with other marine products, such as shells, sea cucumbers, and lobsters (n=4 of 22). A few of the fishers we interviewed at sites near Thailand’s borders   21 reported selling to traders in neighbouring countries for a better price (n=5 of 22; Fig. 3.3). Exchange of commodities (including seahorses) was also reported by one fisher to happen at sea near the Thailand-Malaysia border.   Primary buyers (Level 2 traders) reported that their higher-level buyers (middle-traders, level 3) travelled among coastal cities to gather stock from many primary buyers (n=22), although some received their commodity by post as well (n=3 of 22). The level 2 traders we interviewed reported that while they were loyal to their higher-level buyers, an increasing number of new traders were enquiring about seahorses for sale (n=5 of 22). Level 2 traders reported that they sent the seahorses to where the level 3 traders live (e.g., Ranong, Surat Thani, and Songkhla, as shown in Fig. 3.3), then level 3 traders dispatched a vast majority of their purchased seahorses to Bangkok to consolidators (level 4). From there seahorses were distributed to retailers (level 5) in other regions of Thailand for sale as TCM (n=10 of 22) (Fig. 3.3). Level 2 traders in Phuket also sold seahorses as curios to souvenir stores (level 3; 4 of 9 souvenir stores surveyed in Phuket reported selling seahorses). We could not obtain information about the trade routes from the few level 3 traders (n=3) we interviewed.  Consolidators (level 4) bought seahorses from level 3 traders and sold them to retailers and exporters (n=10). In TCM stores (retailers, level 5), the seahorses were sold by weight, with the unit as gram or “tael” (about 3.75g). Exporters (level 5) obtained most of their seahorses from consolidators (level 4), but sometimes bought them from middle-traders (level 3). Respondents at various trade levels indicated that dried seahorses were exported to China, Hong Kong SAR, and Taiwan as traditional Chinese medicine (n=24) (Fig. 3.3).    22  3.4.2 Trade volume in dried seahorses  Level 2 to 4 traders reported that they sold more seahorses than they bought (Table 3.2). However, exporters (level 5) reported that they bought more seahorses than they sold annually (Table 3.2). Traders from level 1 to 4 reported moving a mean of 17 kg per trawler (Aylesworth et al. under review) up to 297 kg per consolidator. A few exporters (n=3) mentioned that their exports changed after implementation of the Thai export quota in 2012, so we separated their reported exports into 2011 (before the Thai quota) and 2012-13 (when we did our surveys); exporters reported that their mean annual export volume dropped from 1,563 kg in 2011 to 309 kg in 2012-13 (Table 3.2).    Reported domestic consumption of dried seahorses as TCM (from eight TCM retailers) was <10% of the reported export volume. The estimated mean annual sales volume for a TCM store was 0.54±0.52 kg year-1 (n=8), and the estimated total domestic consumption of dried seahorses at around 265±255 kg year-1 (Table 3.2).   3.4.3 Price of seahorses traded in Thailand Mean selling price by trade level ranged from US$0.29±0.11 g-1 (level 1) to US$1.29±0.90 g-1 (level 4) – a four-fold increase (Table 3.3). Mean reported buying price increased 1.5 fold between level 2 and 4 traders (Table 3.3). Retailers (level 5) and exporters (level 5) reported buying and selling prices were lower than those reported by level 4 traders (Table 3.3).     23 The price of seahorses was based on size across all levels of the trade (n=72). Fishers sold seahorses individually, and reported that selling price increased with seahorse size (n=53 of 58 fishers). Higher-level traders (levels 3-5) paid more per kg for a bag of large seahorses than small ones (n=16 of 45); level 4 and 5 traders (n=13) reported the average selling price of the largest seahorses (40-100 seahorse/kg) to be at least four times higher than the average price of the smallest seahorses (300-800 seahorses/kg) (Table 3.3).   3.4.4 Economic scale of dried seahorse trade in Thailand (1) Direct Expenditures: Based on the most recent annual seahorse catch estimate (91±39 tonnes; Aylesworth et al. under review) and mean ex-vessel price (US$0.29 per gram, n=36, Table 3), the direct expenditure (gross output) of dried seahorses in Thailand was around US$26.5 million (95% CI: US$0.1-67.1 million). Within the US$26.5 million, US$18.1 million were generated by commercial fisheries and US$8.4 million came from small-scale fisheries.   (2) Annual domestic consumptive value: The domestic consumptive value was only US$0.2 million (CI: US$0-0.6 million) because of the relatively low trade volume (265±255 kg per year), even though the retail prices in Thailand were high compared to the ex-vessel price (mean=US$0.81±0.25 per gram, n=14, Table 3.3).  (3) Annual gross export values: Annual gross export values were calculated for 2011 and 2013 in order to distinguish the effect of the quota starting in 2012. The annual gross export value for 2011 was US$5.9 million (export =15,256 kg, and mean export price US$0.39 per gram, n=3),   24 and dropped to US$1.0±0.4 millions in 2013 (export=1,430 kg, and mean export price US$0.72±0.26 per gram, n=7).   (4) Number of traders: According to our top-down calculation, the estimated number of traders based on buying and selling volumes (Method 1, ‘B+S’ in Table 2) suggested that the CITES-reported export volume from Thailand in 2011 may involve at least 330 people in the supply chain (not account for level 3 traders or fishers).  In contrast, the CITES-reported export volume from Thailand in 2013 suggested only 23% as many people (again, not account for level 3 traders or fishers) (Table 3.2). If the calculation were based on trade volume (Method 2, ‘T’ in Table 3.2), estimates of the number of traders would decline from around 530 people in 2011 to about 60 people in 2013 (Table 3.2). For both methods and time periods, the estimates for the number of trawlers catching seahorses derived from the export volumes represented 3-33% of the number of trawlers registered in Thai Department of Fisheries (Table 3.2). Using our bottom-up approach, by Method 1, we estimated that an annual catch of about 91 tonnes may involve more than 5,400 people in the supply chain in levels 2, 4, and 5 in 2011 (so excluding number of level 3 traders and fishers; Table 3.2), but only 4,900 people in 2013 (Table 3.2). Method 2 suggested that fewer people were involved in the trade, with 3,400 people in 2011 and 3,100 people in 2013 (Table 3.2).   3.4.5 Income from trading dried seahorses Based on the gross output, we estimated that a commercial vessel could potentially earn US$2,784 from dried seahorses while a small-scale fishing boat could only earn US$227 in a year. If the income from seahorses was evenly divided by the crew as reported to us in some   25 provinces (e.g., Phuket), each crew member could potentially earn US$116-232 p.a. from seahorses (crew size 12-24 per trawler, (DoF, 2015)). However, if based on the total trade volume in 2011, 15,552 kg (export = 15,256 and domestic trade = 265 kg), we estimated that the income from seahorses was US$434 for a commercial fishing boat and US$39 for a small-scale vessel. The decline in export volume in 2013 could potentially reduce the annual per vessel income from seahorse to US$108 for a commercial vessel and just US$8 for a small-scale vessel. Exporters reported their individual mean annual income from seahorses to be US$609,570 in 2011 (1,563,000g × US$0.39) and US$222,480 in 2013 (309,000g × US$0.72).  3.4.6 Perceived changes in trade volume and price Most respondents who described the trend in seahorse trade volume reported a decline trend (8 of 9 primary buyers (level 2), 7 of 7 consolidators (level 4), and 8 of 8 exporters (level 5)). Only one level 2 trader described the trend of dried seahorse volume as stable. Level 3 traders and retailers (level 5) did not report historic trade volumes. The apparent decline in reported volumes over time for primary buyers and exporters was not significant (level 2, slope=-0.13, p=0.10; level 5, slope=-0.21, p=0.07; Figure 3.4a). In contrast, the reported decline in trade volume reported by consolidators was significant (level 4, slope=-0.15, p=0.02, Figure 3.4a); consolidators reported trade volumes in 2013 to be 5% of those they recalled from 20 years ago.   The reported selling price of dried seahorses significantly increased at all levels of trade (Figure 3.4b). The rate of increase in historic selling price was the highest for exporters (level 5, logarithm slope=0.23, p<0.01), and lower with each subsequent descending trade level (slope=0.18 for consolidators, 0.11 for primary buyers, and 0.05 for fishers, all p-value<0.01).   26 It appears that the trend in trade volume did not show a large change after each of the CITES interventions, though exporters have reported that they reduced the exports after Thai government’s suggested quota (Figure 3.4a). Similarly, although respondents at every trade level reported increases in prices, only fishers reported a large increase in selling price corresponding to CITES intervention (Thai quota at 2012).   3.4.7 Comparisons of Thai seahorse exports among different data sources  3.4.7.1 Trade volumes The CITES database reported the dried seahorse export volume from Thailand to range from 9,598 to 20,980 kg per year from 2004 to 2011, with the annual mean estimated at 15,690 kg (SD=3,094). In 2013, the year after Thai government implemented the voluntary annual export quota, CITES data reported a total of just 1,430 kg seahorses exported from Thailand. According to CITES trade records, Hong Kong SAR was the biggest importer of Thai dried seahorses, with an average of 9,896 kg per year from 2005-2013 (SD=4,963), which accounted for 50~90% of the total reported Thai dried seahorse exports (Figure 3.5). Taiwan was reported as the second largest importer overall, but from 2005 to 2013 reported Taiwanese import volumes decreased steadily from 4,791 kg to 318 kg (Figure 3.5). CITES data also reported Thai seahorses being exported to mainland China, Singapore, Australia, New Zealand, USA, Cyprus, and Malaysia from 2004 to 2013.  Hong Kong SAR’s data showed that the mean annual import from Thailand (including seahorses re-exported from Hong Kong SAR) between 1998-2011 was 9,678 kg. The year of the CITES listing, 2002, corresponds to the importing peak of seahorses from Thailand to Hong Kong.   27 Since 2004, Hong Kong’s imports of seahorses from Thailand have gradually declined. In 2013, Hong Kong reported only imported 839 kg of dried seahorses from Thailand, which was the lowest point on record, and lower than reported trade volumes from Thailand to Hong Kong in the CITES database  (Figure 3.5).   Imports from Thailand reported by Taiwan showed a mean annual import of 5,473 kg between 1983-2004, and the quantity started decreasing to a mean annual import of 1,226 from 2005 to 2012 (Figure 3.5a). The imports to Taiwan in 2002 decreased to one of the lowest levels since the records started, but gradually increased until 2004. Then, Taiwan reported that its imports of seahorses from Thailand steadily decreased since 2005, and dropped to only 100 kg in 2012 and 219 in 2013 (Figure 3.5a).  3.4.7.2 Prices Import prices in Hong Kong have been increasing since 2004, and jumped higher in 2012 (Figure 3.6). In contrast, import price in Taiwan has fluctuated with no apparent trend (Figure 3.6). However, the selling prices reported by fishers in Phuket (US$0.06-0.84 per seahorse, or US$0.02-0.27 per gram) were lower than the prices reported by fishers elsewhere.   3.5 Discussion Our study found that the seahorse trade was worth surprising amounts of money and that decreasing volumes in trade (but not in catch) may be responding to CITES-related management interventions. Our quantification of the potential direct expenditures generated by dried seahorses suggested that these bycatch fishes have significant economic importance, though the   28 declared export value was only 20% of calculated direct expenditures. The discrepancy between fisher-reported catch and CITES export volumes suggested that more seahorses might be in the trade than CITES data depicted. Traders perceived a decline in trade volumes over time but did not consider it a result of the trade management interventions. In contrast, however, declared export volume seemed to decline with the global implementation of CITES and with the Thai domestic quota arrived, as has been seen in other wildlife trade (Thorbjarnarson 1999, Raymakers and Hoover 2002).  Thai traders perceived that price increased significantly, and Hong Kong’s import data especially showed an increase, after CITES implementation.     The notable economic value of dried seahorses helps maintain the trade and could be a strong incentive for fishers to continue extracting the fishes from waters. The potential gross output of seahorses found in our study, which we estimated to be US$26.5 millions direct expenditures in a year, was about 1/3 of the declared cash value of total fish for reduction (often dismissed as “trash fish”) in Thailand (US$ 79 million in 2012, OECD 2015). The income from seahorses could be about 20% of the fisheries income of small-scale fisheries (Bennett et al. 2014), and up to 10% of a crew’s earning on commercial vessels (6,500 TBH/month, (International Labour Organization and Asian Research Centre for Migration 2013)). In addition to fishers, exporters could benefit much more by selling seahorses. The mean income from seahorses for an exporter was 37-100 times higher than the mean per capita income in Thailand (US$5,977 for 2011-2016, (The World Bank 2016)). Such estimate of exporters’ income might even be underestimated, since the export prices reported by exporters were lower than the selling prices reported by consolidators. Given the apparent cash value of seahorses, there may need to be other incentives to support conservation actions for seahorses (Roe et al. 2002) particularly   29 given that the seahorses are bycatch (Yasué et al. 2015). The economic revenue from bycatch can help subside the fishers to continue fishing even if the target species are over-exploited (Branch et al. 2013).   The huge mismatch between the Thai seahorse catch estimate (Aylesworth et al. under review) and CITES declared export volume suggests that a large part of the trade might be unreported, creating a concern for effective management (Srikosamatara et al. 1992, Shepherd and Nijman 2007). In our top-down calculations for estimating the number of traders, both methods agreed that the number of trawlers deduced from declared export volume was far lower than the actual number of registered vessels (DoF 2015). This discrepancy could arise if most seahorse catch is being retained rather than traded. Given the high economic values of seahorses, however it is more likely that a significant number of seahorses in trade went unreported in the CITES records (Blundell and Mascia 2005).  We could not validate the CITES database independently since the exporters we interviewed had all registered for CITES permits, and scaling up from lower levels of trade requires full knowledge of how many traders are operating at these levels. The large gap between trade records and catch made the declared export volumes an inaccurate indicator of extraction levels, and might create a delay in noticing and addressing over-exploitation (Blundell and Mascia, 2005; Cooney and Jepson, 2006; Foster et al., 2016).   It seems that fisheries catch of seahorses continued untrammelled even when there are hints of changing exports after CITES interventions.  Trade volumes of primary buyers were not strongly affected by CITES interventions even though Hong Kong and Taiwan’s data showed that Thai exports of seahorses had declined after both global implementation of CITES and the domestic   30 restrictions on exports (Beissinger and Bucher. 1992, Lee and Smith 2010). The declining trend of seahorse exports after implementing CITES was consistent with the trade of other CITES Appendix II species; for example, Kenya’s exports of chameleons declined following the adoption of national management (Carpenter 2004). However, our respondents at lower trade levels did not report that trade volumes were affected by CITES implementations and domestic regulations. Such inconsistent, or delayed responses in the changes in trade volume among trade levels have also been found in other wildlife trade when export quotas were applied (Beissinger and Bucher. 1992, Lee and Smith 2010). In our case, it probably indicates that reducing exports has limited effects on bycatch extraction of fishes or, therefore, on lower levels of trade (Casale 2010). Moreover, the changes in declared exports might not be the true export volume, as a certain proportion of trade might be unreported (Blundell and Mascia, 2005; Foster et al., 2016).    The increases in price of trade-regulated seahorses, as seen in every trade level in our study, might result from the rising international demand, coupled with declining supply because of overfishing or/and trade restrictions (Challender et al. 2015a).  The majority of trawl fishers in Thailand reported perceived declines in seahorse catch (Aylesworth et al. under review) even while exporters and wholesalers we surveyed indicated that the demand from Mainland China and Hong Kong had been increasing. In addition, after CITES implemented its listing of seahorses, many large source countries, such as Philippines, banned export of seahorses (Yasué et al. 2015). Facing the significant reduction in seahorse supply, the international market might depend more on the few exporting countries, including Thailand. That was not surprising that the import price of seahorses to Hong Kong would increase almost 7 times from 2004 to 2011. However, the import prices of Taiwan have fluctuated and did not show this apparent trend,   31 suggesting a different market response to Hong Kong. Although seahorses were mainly bycatch in Thailand, the rise in the price still needs to be carefully considered as a driving force for further exploitation. Taking the example of sharks, while only a few pelagic shark species (e.g., Isurus oxyrinchus and Lamna nasus) were targeted in the past for the meat (Dulvy et al. 2008), the increasing value of shark fin has led to more shark species (e.g., Prionace glauca) being targeted (Dulvy et al. 2008).  This study highlights one of the challenges that arises in implementing CITES Appendix II listing of species that are obtained in by-take or by-catch. Even if countries seek to manage exports in conformity with their CITES obligations by imposing an export quota or some other restriction, this may only result in reduced exports at best and not in reduced catches or impacts on the wild populations. That said, even Thailand has suspended its seahorse exports in 2016, improving the fisheries management to reduce bycatch is more critical for seahorses. It may generally be much more effective in conservation terms for Parties to eschew the allure of a simple fix through export measures, and instead to manage extraction in a way that allows for exports. One of CITES’ recommendations for Thailand in the 2012 was that Thailand increases its enforcement of fisheries restrictions (e.g., the 3 km wide trawl exclusion zone along the coast of the entire country) (CITES AC26 2012). The current concerted effort from Thailand to address Illegal, Unregulated and Unreported (IUU) fishing – goaded by European Union concerns (Leroy et al. 2016) - will advance national capacity to improve the trawl fisheries.  The goal of CITES Appendix II listings is, after all, to ensure persistence of carefully managed trade, not to close it down.    32 CITES is a promising tool for ensuring the sustainable trade of marine species, but it is not a stand-alone solution. CITES has urged governments to improve their management for the protected species, and successful stories have been found in terrestrial species (Aiken et al. 1999, Sadovy et al. 2007, Recharte Uscamaita and Bodmer 2010a). However, reforming international/national management and enhancing enforcement is still the key to controlling exploitation. Given that small bycatch fishes comprise 30-40% of the total catch from Thai waters (Ahmed et al. 2007), managing seahorse extraction and ensuring the sustainability of its trade has implications for the management of other small bycatch species. Bycatch species can be highly valuable – as our case shows – and ensuring the sustainability of their trade would be not only beneficial to the species, but also to the people depending on them.  33 Table 3.1 Number of respondents interviewed during trade surveys in Thailand, categorized by occupation and location.        Occupation Eastern and central Gulf of Thailand Southern Gulf of Thailand Andaman Sea Total Fishers (Level 1) 47 24 27 98 Primary buyers (Level 2) 8 8 6 22 Middle traders (Level 3) 3 0 0 3 Consolidators (Level 3, 4) 5 0 5 10 Internet traders (Level 2, 3, 4) 9 1 0 10 Souvenir stores  0 0 10 10 TCM retailers  16 1 6 23 Exporters 9 0 0 9 Others (Gov. officers, NGOs, residents…etc.) 7 7 4 18 Total 104 41 58 203   34 Table 3.2 Trade volumes and number of traders of dried seahorses for each trade level in Thailand. Data obtained from external datasets or literature are in italic. We used reported mean catch from otter trawlers to representative fishers’ trade volume (level 1). Number of traders was estimated based on the mean trade volume and by two methods (See Appendix B) for sensitivity analysis: (1) considering both reported purchase volume and sell volume (P+S), and (2) considering only the maximum trade volume (T).  Top-down calculation Bottom-up calculation CITES 2011 CITES 2013 Catch estimate§ (2015) Volume (kg) 15,256 (UNEP-WCMC 2013a) 1,430 (UNEP-WCMC 2013a) 91,677±77,220§ Annual domestic consumption volume (kg) 265  Mean volume per trader ±95%CI¶  (kg trader-1 year-1)  Estimated number of traders  Purchase Sell Trade P+S T P+S T P+S T Exporters – 2011 (Level 5) 1,740  (n=1) 1,475  (n=2) 1,563  (n=3) 10 10 -- -- 155 59 Exporters –2013 (Level 5) 391±364  (n=5) 157±60  (n=7) 309±235  (n=8) -- -- 9 5 692  297   35  Top-down calculation Bottom-up calculation CITES 2011 CITES 2013 Catch estimate§ (2015) Consolidators (Level 4) 242±141  (n=6) 341±323  (n=4) 297±150  (n=10) 53 52 11 6 794 309 Middle-traders (Level 3) 30  (n=2) 360  (n=1) 140  (n=3) -- -- -- -- -- -- Primary buyers (Level 2) 23±12  (n=18) 48±60  (n=8) 33±23  (n=21) 267 466 56 51 3,967 2,776 Total # of traders (Level 2, 4, 5) Year: 2011 330 528   4,916 3,144 Year: 2012   76 62 5,453 3,382 Trawlers (Level 1)  17±17  (n=38)§ 361 908 78 100 -- # of otter trawlers registered in Thai Department of Fisheries 2553 (DoF 2015) ¶ Confidence interval is reported when sample size is larger than 3. § Data reported in Aylesworth et al. under review    36 Table 3.3 Mean purchasing and selling price of dried seahorses in each trade level (USD per gram) across three seahorse size categories (Small (s), Medium (M), Large (L)). Sample sizes (number of respondents) are shown next to the prices in brackets. Purchasing price in USD per gram Selling price in USD per gram  Mean ± 95% CI (n) Mean price by size (n) Mean ± 95% CI (n) Mean price by size (n)   S M L  S M L Exporter – 2012, 2013 (level 5) 0.71±0.26 (6) 0.25 (2) 0.56 (4) 0.97 (4) 0.72±0.26 (7) 0.13 (3) 0.46 (4) 1.15 (5) Exporter – 2011  (level 5)     0.39 (3)    Retailer (level 5) 0.33±0.2 (4)    0.81±0.25 (14)    Consolidators  (level 4) 0.81±0.27 (7) 0.39 (4) 0.93 (3) 1.70 (2) 1.29±0.90 (9) 0.31 (8) 0.73 (7) 1.11 (6) Middle-traders  (level 3) 0.30 (3) 0.32 (1) 0.49 (1)  0.44 (2) 0.48 (1) 0.58 (2)  Primary buyers  (level 2) 0.53±0.48 (9)    0.68±0.41 (21)  0.35 (3) 0.37 (1) Fishers (level 1)  0.29±0.11 (36)  0.42 (2)    37  Figure 3.1 Provinces in Thailand surveyed during the trade surveys in 2013-2014.     38  Figure 3.2 Dried seahorse trade structure in Thailand. Arrows indicate the direction of trade flow, from oceans to export.  39  Figure 3.3 Potential trade routes for dried seahorses in Thailand as deduced from trade interviews. The arrows show the direction of trade flows. Survey locations are indicated as solid circles, whereas open circles represent places identified by respondents.  40  Figure 3.4 Changes in (a-c) mean trade volume and (d-g) per-gram selling price of dried seahorse reported by traders in each trade level in Thailand. Data were log transformed. The changes in trade volume and prices are presented by linear regression lines   41 (if significant). The three interventions are marked by dashed lines: (i) CITES listing, (ii) CITES implementation, (iii) Thailand voluntary export quota implemented.   42  Figure 3.5 A comparison of international trade volume from Thailand using three official datasets (CITES, Hong Kong CSD, and Taiwan Customs). Records from importers were stacked into one bar for each year; Hong Kong Census and Statistics data were in dark grey and Taiwan customs data were in light grey. For the CITES data, trade volume from Thailand to Hong Kong (HK) and Taiwan (TW) were shown in the same colour code as the Customs data but with stripes (HK: dark grey; TW: light grey), and to other destinations were in black. The three interventions were marked by arrows: (i) CITES listing, (ii) CITES implementation, (iii) Thailand voluntary export quota implemented.   43  Figure 3.6 Comparing prices that Hong Kong (HK) and Taiwan (TW) imported seahorses from Thailand. The three interventions are marked by arrows: (i) CITES listing, (ii) CITES implementation, (iii) Thailand voluntary export quota implemented.   44 Chapter 4: Assessing the impacts of CITES implementation on the international trade of marine species – a case study of seahorses  4.1 Synopsis Trade regulations may be useful for conserving marine species that are suffering from overexploitation. The Convention of International Trade on Endangered Species of Flora and Fauna (CITES) has emerged as an instrument to help tighten fisheries management. However, the impacts of CITES regulations have not been examined for the trade in fully marine fishes. In this study, we used seahorses (Hippocampus spp), the first fully marine fishes listed in CITES Appendix II since treaty inception, as a case study. Drawing on Customs data from Taiwan and Hong Kong SAR (which cover pre-CITES periods), we applied iterative segmented regression to investigate changes in seahorse trade corresponding to CITES interventions. We then conducted principal component analyses to understand characteristics of seahorse source countries, and applied a gravity model of trade to identify predictors of seahorse trade volumes. We found that the total weight of seahorses in documented trade decreased significantly after CITES implementation, recorded trade became concentrated in fewer countries, and prices increased. We also found that seahorse source countries had more fishers, demersal fish catch and general trade with China, compared to other range states. However, countries that reported no exports, unchanged export volumes or declining volumes after CITES were similar. In addition, we found that volumes traded between two countries were significantly higher when they were closer together or when the source country had a lower per capita GDP or higher demersal catch. Our study can help guide targeted actions to maximize CITES effectiveness for marine species.   45 4.2 Introduction Commoditization of wildlife has shifted the driver of wildlife exploitation away from supporting livelihoods toward supporting local and global markets (FAO 2016), and increased the scale of exploitation to an extent that it poses a significant threat to species survival. Over-exploitation from activities such as logging, hunting, and fishing, directly leads to population declines and habitat destruction as evidenced by the more than 2,700 animal species listed as near-threatened or threatened on the IUCN Red List (Maxwell et al. 2016). Wildlife is extracted not only for subsistence use, but also for local and international timber, food, medicine, fashion and pet markets, among many others. The global value of imports of wildlife products was estimated in 2009 at about USD 323 billion, coming from trade in tens of thousands of species (Engler 2008). Monetizing the value of species could accelerate the exploitation rates of wild animals and plants (Courchamp et al. 2006). For example, despite a long history of local consumption of Sunda pangolins (Manis javanica) and Burmese starred tortoise (Geochelone platynota), increasing international market demands have driven these species to near extinction in just a few decades (both species are now considered Critically Endangered on the IUCN Red List).    When trade expands to global levels, its large-scale and asymmetric nature result in disproportionate exploitation among different regions, and difficulties in management (Crona et al. 2015, Challender et al. 2015a). Globalization allows consumers to access natural resources across borders. When demand increases beyond what a country can provide, buyers seek new suppliers leading markets to expand to other countries (Berkes 2006). Such exploitation expansion has been well documented in the trade of many species, including sea cucumber (Anderson et al. 2011) and sea urchin (Andrew et al. 2003). Following the expansion in trade,   46 serial depletion has been identified in several local resources (e.g., (Orensanz et al. 1998)). Industrialized countries consume an unequal proportion of traded environmental resources when compared to less developed countries (Watkins and Fowler 2002). This typically leads to over exploitation of natural resources in developing countries because they often lack capacity and resources to manage such exploitation (Chichilnisky 1993). While the supply and demand of international trade are separated from local management efforts, collaborations among national governments were urgently required to conserve global biodiversity.  The largest collaboration for regulating the complicated international wildlife trade is the Convention of International Trade on Endangered Species of Flora and Fauna (CITES) (Vincent et al. 2013). The main purpose of CITES is to ensure the sustainability of wildlife under the globalization of trade (CITES 1973). Species listed in different CITES Appendices are according to the degree that they are threatened by trade (CITES 1973). Trade in Appendix I species is basically prohibited, while the trade of species in Appendix II has to be accompanied with permits and determined not detrimental to the wild populations. Appendix III includes species that one range state (countries where the species occurs) has asked other countries to assist in protecting their sustainability. To respond to CITES’ requirements, countries have used various methods to control their wildlife supply, including limiting the number of hunting licenses, closed seasons, and bans (Barden et al. 2000, Raymakers and Hoover 2002). However, if and how those efforts lead to changes in trade are still unclear.   To date, evaluations of changes in wildlife trade volume and prices linked to CITES listings have varied across countries and species. Multiple factors have been associated with changes in trade   47 under CITES, such as source countries’ capacity in improving management and historic/cultural value of the trade to stakeholders (Di Minin et al. 2015, Robinson et al. 2015). For example, for the amphibian and reptile species listed in CITES Appendix II, their trade volume of wild animals has declined globally (Nijman and Shepherd 2010, Robinson et al. 2015). This was due in part to the success of ranching and captive-breeding activities (Nijman and Shepherd 2010, Robinson et al. 2015). Such increases in captive-bred animals mostly happened in the countries where funding and expertise are available (Herrel and Meijden 2014, Robinson et al. 2015). In addition, increased retail prices have also been documented for various CITES species, including mammals, amphibians, and reptiles (Courchamp et al. 2006, Challender et al. 2015a). However, previous studies about the impacts of CITES regulations on trade have mainly focused on terrestrial species. Currently we have little understanding of CITES’ impacts on marine fishes – one of the biggest groups that suffer from over-exploitation.   Can CITES also provoke changes in the trade of marine fishes? If so, what are the determinants of such changes? In 2014, more than 78% of the seafood products, in which 81 millions tonnes were wild-caught marine fishes, went into international trade (FAO 2016). Analysis of FAO trade data for fisheries products indicated that bilateral trade volume was determined by geographical distance between two countries, the production volume of the source country, per capita consumption in the destination country, and regional trade agreements (Natale et al. 2015). But when there is a trade restriction, do these factors still associate with the variations in trade volume? Previous analyses of the effects of food safety standards set by the United States, Japan and European Union found such restrictions in seafood imports benefit the developed countries than the developing counterparts (Anders and Caswell 2009, Baylis et al. 2011). Trade   48 regulations for combatting illegal fishing, e.g., EU’s “yellow cards”, were not found to have impacts on the sources of seafood imports, at least not when the analysis was done in 2014 (DG MARE 2014). However, trade sanctions on selected countries issued by regional fisheries management organizations (e.g., ICCAT) have resulted in decreasing imports of the regulated fish species from those states (Hosch 2016). Since 2002, an increasing number of marine fishes have been proposed for CITES listing (Vincent et al. 2013, Cochrane 2015). However, the impact of CITES on the global trade patterns in marine fishes have not been studied.   Seahorses (Hippocampus spp.), the first marine fishes listed in CITES Appendix II since its inception, serve as an invaluable example to examine the impacts of CITES on trade. Seahorses are mainly traded dry for traditional medicine and curios, but also live for aquarium uses. Around 37 million dried seahorses are caught incidentally by non-selective gears each year, and the trade is widely occurring across the globe with as many as 80 countries involved (Vincent et al. 2013, Foster et al. 2016, Lawson et al. 2017). All seahorse species were listed in CITES in 2002, and the listing was implemented in mid 2004. Customs records from two major seahorse markets, Hong Kong and Taiwan, contain seahorse import data from the pre-CITES period (1983 and 1998, respectively) and provide an opportunity to investigate the changes in seahorse trade regarding to CITES intervention.   Our study aims to investigate if and how CITES affects the trade volume, prices, and trade routes of marine fishes, using seahorses as a case study. Moreover, we explore if the different country-level characteristics across countries can be indicators for the changes in each country’s seahorse exports. We focus on the trade of wild, dried seahorses, which accounted for 97% of all   49 seahorses reported in trade (Foster et al. 2016). We hypothesize that the global trade volume of seahorses would decline and the supply of seahorses would be dominated by few countries after CITES listing, because countries that were unable to ensure the sustainability of their trade had to suspend their exports. In contrast, we hypothesize that the prices would increase since the demand was not satisfied by the decreased volumes.   4.3 Methods To identify the changes in trade after CITES interventions, we collected seahorse trade data from two Customs datasets and the CITES trade database. We also explored if country attributes associated with (1) whether a range state reported as exporting seahorses; (2) whether exports from a seahorse source country were reported as significant dropping or stopping after CITES implementation; and (3) the bilateral trade volume. We describe our datasets and analyses as follows.   4.3.1 Trade data We used three independent data sets to examine the changes in global seahorse trade over time: (1) import and re-export data from Hong Kong Census and Statistics Department (CSD) (Hong Kong Census and Statistics Department 2015), (2) import data from Taiwan Customs (https://portal.sw.nat.gov.tw/APGA/GA03, accessed October 31, 2015), and (3) the CITES trade database (http://trade.cites.org, accessed June 24, 2016). While the Hong Kong and Taiwan’s data include both quantity and price of dried seahorses, CITES data included only the quantity traded.     50 We sourced data on the quantity and prices of global seahorse trade from Hong Kong Census and Statistics Department for 1998-2014 (CSD, Hong Kong Department of Census and Statistics 2015). We analyzed imports from Country of Origin - countries where products were produced or had undergone the last permanent transformation (Hong Kong Census and Statistics Department 2015). We did not analyze Country of Consignment, the products’ last stop before Hong Kong, because they were usually not the source of seahorses (Hong Kong Census and Statistics Department 2015). We converted the import prices in CSD statistics from Hong Kong Dollars to US Dollars, based on the exchange rate of each year, (http://www.usforex.com, accessed June 20, 2016,). Note that Hong Kong CSD data are independent of Hong Kong’s reports to CITES, since CSD are Customs records and CITES reporting in Hong Kong is controlled by the Agriculture, Fisheries and Conservation Department (AFCD). The reports from AFCD are based on the CITES permits submitted by the importers and re-exporters.  Seahorse trade data from Taiwan’s Customs covers 1983-2014 (https://portal.sw.nat.gov.tw/APGA/GA03, accessed October 31, 2015). Taiwan’s data include the annual dried seahorse import weights from the origin and import values (in USD), however the import values in the dataset were only broken down by country after 2002. We divided the import values by the import weights to calculate the mean price per kilogram of seahorses from each country. Since Taiwan is not a member of CITES, the Customs data is independent to the data in CITES database. The data of seahorse traded to or from Taiwan in the CITES database are reported by CITES members.    51 From the CITES trade database (http://trade.cites.org, accessed June 24, 2016) we extracted information from 2005-2013 to examine the global pattern of seahorse trade in the post-CITES period. We excluded CITES data from before 2005, since CITES implementation for seahorses started in May 2004 and prior to this countries were not required to report their trade. We considered commercial trade of all Hippocampus species and only analyzed the trade volume of wild dried seahorses (97% of total dried trade). In the database, we considered records with the terms ‘bodies’, ‘derivatives’, ‘specimens’ and ‘skeletons’ were dried seahorses (Foster et al. 2016). All analyses for CITES data were performed on the Exports from imports data (EFI, defined in Foster et al. 2016), which were the larger volume of the reports from exporters and importers when they did not match. For each data entry, if there was no unit reported, we assumed the unit was individual per UNEP-WCMC guidelines, except for the ones clarified by the reporting countries (Foster et al. 2016).   4.3.2 Analysis 4.3.2.1 Changes in declared seahorse trade after CITES listing We used Hong Kong CSD and Taiwan Customs data, which cover pre- and post-CITES listing periods, to examine the following four changes in declared seahorse trades corresponding to CITES interventions: 1) total weight of seahorses in trade, 2) number of source countries, 3) evenness in the supply of seahorses among source countries, and 4) import prices. The two CITES interventions we were interested are: the listing of Hippocampus sp. in Appendix II in 2002, and implementation the listing in 2004.     52 We conducted iterative segmented regression (Toms and Lesperance 2003) for each trade variable (total weight of seahorses in trade, number of source countries, evenness in the supply and import prices), to test if there was a sharp change in the data from a certain time point, and whether the breakpoint corresponded to CITES interventions. Our model for the regression is:     𝑉𝑡 = 𝑎0 + 𝑎1𝑌𝑒𝑎𝑟𝑡 + 𝑎2𝑌𝑒𝑎𝑟𝑡 ∙ 𝐼 + 𝑎3𝐼    (Eq. 4.1) For which we regressed the trade variable (V) with time (Year). Dummy variable I is 1 if Year is smaller or equal to the breakpoint of interest, and 0 if Year > the breakpoint. Because of the sample size restrictions, we tested breakpoints from 2003 to 2007 for Hong Kong’s data, and 2000 to 2007 for Taiwan’s data. We iteratively searched the “best” breakpoint for the model that has the highest r-square value. We hypothesized that the “best” breakpoint would be at 2005, the year after CITES implementation.    We calculated the Gini Index as a measure of the evenness of seahorse supply (Gini 1921). Supply is more uneven if the Gini Index is closer to one, and is more even if the index is closer to zero. We calculated the Gini index over time for each data set separately, and considered all source countries that existed in that dataset. Although we only applied the segmented regression on Hong Kong and Taiwan’s data, we still calculated the Gini index for CITES for comparison. For CITES data, the Gini index was calculated for all source countries regardless of export destination.   4.3.2.2 Linking country-level characteristics to changes in seahorse exports We explored three main questions to understand how country-level characteristics may correlate with reported seahorse exports. First, for seahorse range states – countries with one or more   53 seahorse species found within their exclusive economic zone (UNEP-WCMC 2015) - we investigated why some range states were documented seahorse exporters, while others were not. Second, we examined the differences in characteristics of source countries that purportedly continued or stopped exports after CITES implementation. Third, we used a gravity model for trade to understand if and how bilateral trade volumes correlated with the characteristics of reported source and destination countries.   4.3.2.2.1 Why some seahorse range states were documented seahorse exporters, while others were not?  We conducted a principal component analysis (PCA) to determine the key characteristics that differentiate the range states reported as seahorse sources and those that are not. Seahorse sources were sovereign states reported as exporting seahorses in any of the Customs or CITES datasets (1983-2014).   We considered the following five country characteristics in our PCA (see also Appendix C1): 1) demersal fish catch volume by the country’s vessels; 2) distance from China (the main market of dried seahorses); 3) the annual value of a country’s trade in general goods with China; 4) number of fishers working in marine fisheries (including commercial and artisanal fishers); and 5) per capita GDP. We used 1) demersal fish catch to represent potential seahorse catch for each country. Adult seahorses are demersal and mainly obtained as bycatch (Lawson et al. 2017), so we predicted that demersal fish catches might be positively correlated with seahorse bycatch. China is one of the most important market for dried seahorses, and so we used 2) the distance   54 between China and the reported source country, together with 3) their bilateral trade in other goods, as indicators of the country’s accessibility to the Chinese market. Our calculation of the distances between China and another country weighted the populations of the top 25 cities in each country (Head and Mayer 2010). We used 4) the number of fishers in marine fisheries as an indicator of the fishing effort. Finally, we used 5) the average per capita GDP to represent the wealth of each country.   To align with the time series of seahorse trade data, we aimed to collect attribute data from 1983-2014. This was only possible, however, for per capita GDP. For demersal catch and trade with China we could only obtain data from 1983-2013 and 2005-2014, respectively. We took the mean of the time series for the PCA. Each country’s marine fisheries employment was only estimated for 2003 (Teh and Sumaila 2013). The sources of data used to determine country characteristics, and details of each attribute are described in Supplemental Information A.   4.3.2.2.2 Why do some source countries purportedly continue their exports after CITES implementation, while others reported stop or having significant decline in exporting?  We conducted a principal component analysis to investigate if countries that reportedly stopped exporting seahorses after CITES implementation have different characteristics than the countries that supposedly continue trade. We combined Hong Kong CSD, Taiwan Customs and CITES data to determine all countries with a reported history of commercial trade in dried seahorses (n=33). We considered source countries to have “stopped” exports if they had no records of exporting seahorses post-CITES in any of the three datasets, or had a dramatic drop in reported   55 trade (to <10% of mean annual reported volumes in pre-CITES period) in reported imports to Hong Kong or Taiwan (n=22). The remaining countries (n=11) we considered to have continued exporting seahorses after CITES implementation.   We explored three hypotheses related to the country-level characteristics of a change in behavior with CITES implementation. First, we hypothesized that a country would stop issuing export permits for seahorses if they were unable to ensure sustainable trade. In such cases, there may not be enough capacity and/or funding to manage their national fisheries. Second, we hypothesized that countries would have more motivation to continue the trade if there were more people involved in the national fisheries. Third, we hypothesized that a country may continue to export seahorses if there was historically high seahorse export volume.    In our PCA, we explored the following six variables we believe would be linked to a change in behavior after CITES implementation: 1) demersal fish catch, 2) per capita GDP, 3) beneficial subsides to sustainable fisheries (Sumaila 2010), 4) fisheries capacity-building subsides (Sumaila 2010), 5) number of fishers in marine fisheries, and 6) seahorse export volume before CITES implementation (using Hong Kong imports as representatives). Because seahorses are mainly bycatch in demersal fisheries, we hypothesized that the 1) demersal fish catch would indicate the level of the demersal fisheries requiring management. We considered that 2) per capita GDP, 3) beneficial subsides, and 4) fisheries capacity-building subsides may be indicators for a country’s financial capacity for implementing CITES. We separated the 3) beneficial subsides (investments in managing sustainable fisheries) from 4) fisheries capacity-building subsides (for building a bigger boat, fuel supply, etc.), because they may indicate the different attitude of a government   56 toward managing its fisheries (Sumaila 2010). The number of fishers in marine fisheries 5) represented the number of fishers involved in the fisheries, and 6) pre-CITES seahorse exports represented the importance of seahorse trade to a country.   4.3.2.2.3 What determines the reported trade volume between two countries before and after CITES implementation?   To identify the predictive variables for seahorse trade volume, we fitted gravity trade models (Dascal et al. 2002, Natale et al. 2015) to the Hong Kong’s CSD imports, Taiwan Customs, and CITES data. For the Hong Kong and Taiwan’s Customs data, the model was applied on pre- and post-CITES data separately. Despite recognizing that the declared trade volume might be under-reported, we assumed that the declared volumes were proportional to the real volumes and represented the relative trade volumes among countries. In our analysis, we excluded the reported trade that was from unknown sources.   The gravity model of trade describes the bilateral trade volume as a logarithm relationship with country characteristics. For the trade volume between two countries i and j (xij), we tested the following eight predictors: the geographical distance between two countries (Distij); GDP per capita of source (GDPPCi) and destination (GDPPCj); number of fishers in the marine sector of each source country (Fpopi and Fpopj); demersal fish catch (Dcatchi and Dcatchj); and year (Yeart).   The model is structured as below:   57 ln 𝑥𝑖𝑗𝑡 = 𝛼1 + 𝛼2 ln 𝐷𝑖𝑠𝑡𝑖𝑗 + 𝛼3 ln 𝐹𝑝𝑜𝑝𝑖𝑡 + 𝛼4 ln 𝐹𝑝𝑜𝑝𝑗𝑡 + 𝛼5 ln 𝐺𝐷𝑃𝑃𝐶𝑖𝑡 +𝛼6 ln 𝐺𝐷𝑃𝑃𝐶𝑗𝑡+ 𝛼7 ln 𝐷𝑐𝑎𝑡𝑐ℎ𝑖𝑡 + 𝛼8 ln 𝐷𝑐𝑎𝑡𝑐ℎ𝑗𝑡 + 𝛼9 𝑌𝑒𝑎𝑟𝑡 (Eq. 4.2) Where ijt represents the value of a variable for source i and destination j in a given year t.   We had to overcome two challenges with using a gravity model in order to analyze our data. First we needed to address that trade between two countries may also be influenced by trade interactions among other countries (Baier and Bergstrand 2007). This is commonly referred to as multilateral resistance term (MRT). We therefore added a MRT using Baier-Bergstrand first-order Taylor-series approximation (Baier and Bergstrand 2007) to account for the multilateral resistance for the explanatory variable “distances”. The second challenge we overcame was that gravity models typically could not account for zeros because they were logarithmic. To account for this, we added a small value (10-10) to the zeros to incorporate small trade volumes that were hardly detected (Linneman 1966).   4.4 Results 4.4.1 Changes in declared seahorse trade after CITES  Hong Kong CSD and Taiwan Customs data both show that declared weight of dried seahorse imports to the two major markets have declined in the post-CITES period. For Hong Kong, the declining in import weights since 2005, the year after CITES implementation. In contrast, the imports to Taiwan had the highest change in trend at 2001, suggesting the imports started decreasing before CITES listing (Table 4.1 & Figure 4.1). Based on Hong Kong CSD data, mean annual imports decreased from 21.9 tonnes of dried seahorses during the pre-CITES period   58 (1998-2004) to 7.1 tonnes per year (2005-2014). Similarly, imports also decreased when we analyzed the Taiwan Customs data from a mean of 1.3 tonnes annually imports in pre-CITES period (1983-2014), to only 0.9 tonnes after CITES implementation. CITES records, reported by exporters or importers, were consistent with the trend observed using Customs data (Figure 4.1).   Hong Kong CSD showed that the number of declared source countries decreased after CITES implementation, while the trend in Taiwan Customs data did not significantly change (Table 4.1 & Figure 4.2). Hong Kong reported sourcing from a mean of 13 countries in each year during pre-CITES period to 5 in post-CITES period (Figure 4.2). Taiwan’s imports were reported sourcing from a mean of 6 countries p.a. before CITES implementation, but only reported sourcing from 2 countries for each year after 2005 (Figure 4.2). In addition to the declines in the number of source countries, the seahorse supply became concentrated in very few countries (Table 4.1 & Figure 4.3). The evenness in supply to Hong Kong has slightly declined but that trend started before the CITES implementation, with no change in slope over time (Table 4.1 & Figure 4.3). The evenness in supply of Taiwan’s imports declined significantly faster after 2002 (increases in Gini Index) (Table 4.1 & Figure 4.3).  The composition of source countries for both Hong Kong and Taiwan’s changed with the CITES listing and implementation (Figure 4.4). Thailand became a more dominant supplier of dried seahorses after 2004 (Figure 4.4 & 4.5). Moreover, we found an expansion in source countries from Asian to African countries in both Hong Kong and Taiwan over time (Figure 4.4). For Hong Kong’s data, we found that not only the composition and share changed in the source   59 countries, but also in its re-export destinations, as the re-exports to mainland China dropped greatly after CITES listing (Figure 4.5).   Import prices in Hong Kong have increased significantly since CITES implementation (Table 4.1 & Figure 4.6). The prices of Taiwan’s imports have also increased over time, although we did not have a sufficient sample size to test the CITES listing and implementation (Figure 4.6). The rate of price increases were higher in Asian source countries than in other regions, especially Africa (Figure 4.6).  4.4.2 Linking country-level characteristics to changes in seahorse exports 4.4.2.1 Why some seahorse range states were documented seahorse exporters, while others were not? The distribution of seahorse source countries and non-source countries overlapped a large proportion on the PCA map, suggesting sharing similar countries characteristics (Figure 4.7). However, source countries generally have higher demersal catch, more people employed in marine fisheries, shorter distances from China and more trade with China, compared to the seahorse non-source countries (Figure 4.7).   4.4.2.2 Why do some countries purportedly continue their exports after CITES implementation, while others stop or having significant decline in exporting? Countries that continued issuing seahorse export permits after CITES tended to have lower per capita GDP and higher pre-CITES trade, however their country attributes were generally similar to other source countries (Figure 4.8).    60  4.4.2.3 What determines the reported trade volume between two countries before and after CITES implementation?  We found that the results of gravity model were consistent between the pre-CITES and post-CITES periods for Hong Kong’s imports. However, while the model identified two predictors (number of fishers and distance) for Taiwan’s pre-CITES imports, no predictors were significant for Taiwan’s imports in the post-CITES period.   In Hong Kong’s Customs data, for both pre-CITES and post-CITES period, more seahorses reportedly arrived from the source countries with fewer fishers (p<0.01), lower per capita GDP (p<0.01), higher demersal fish catch (pre-CITES: p=0.03; post-CITES: p<0.01), and greater distance from Hong Kong (pre-CITES: p=0.02; post-CITES: p<0.01) (Table 4.1).   In Taiwan Customs data, more seahorses were putatively imported from source countries with more fishers (p=0.02) and farther from Taiwan (p<0.01) in the pre-CITES period. No predictor was found significant for Taiwan’s imports post-CITES.   For all pairs of countries trading seahorses post-CITES, reported volumes were higher from source countries with lower per capita GDP (p=0.02) and higher demersal fish catch (p<0.01), to destination countries that also had higher demersal catch (p<0.01). The analysis of CITES data also showed that countries with shorter distance between each other recorded traded more seahorses (p<0.01). Fewer seahorses were recorded traded over the years (p<0.01).     61 4.5 Discussion In this first analysis of the effect of CITES listing on marine fishes, we found changes in various aspects of global seahorse trade in the post-CITES period. CITES listing and implementation corresponded with declines in the documented trade volume of wild-caught seahorses, concentrations in supply, and increases in recorded prices, similar to the findings in the live reptile trade (Robinson et al. 2015). We also found that range states that ever reported exporting seahorses generally had larger demersal fisheries catch, highlighting the importance in managing such non-selective fisheries for sustainable seahorse trade (Lawson et al. 2017). However, those countries continuing the trade after CITES implementation did not always invest greater efforts in ensuring sustainable fisheries. In addition, our findings that bilateral seahorse trade volume correlates with geographic distance and the scale of marine fisheries and economics was consistent with other studies on seafood trade (Natale et al. 2015). These results provide insights in examining the impacts of CITES, predicting changes in seahorse trade volumes, and identifying potential under-reporting (Bruckner 2001, Pernetta 2009, Poole and Shepherd 2016).   In our analyses, CITES interventions correlated with steeper declines in declared trade volume of wild seahorses, consistent with the trend found in other CITES listed species (Wood et al. 2012, Robinson et al. 2015). The more rapid decline in recorded trade volume after listing could have arisen from three national responses to the CITES implementation: (1) some countries managed to control properly their export trade in wild seahorses, (2) some countries officially suspended seahorse exports and (3) some trade proceeded without permits, disappearing from formal statistics:  (1) Some countries managed to control properly their export trade in wild seahorses   62 The first possible response, Parties meeting their CITES obligations, was unlikely for the exporters that are big enough to affect global trade statistics. In particular, Thailand, which generated 60% of Hong Kong’s imports and 70% of Taiwan’s before CITES listing, was repeatedly given formal recommendations by CITES because it failed to meet its obligation to the Convention (CITES Secretariat 2016).    (2) Some countries officially suspended seahorse exports In accordance with the proposed second response, country suspensions, some other big traders ceased exporting after CITES listing, either because of national rules or because of difficulties in meeting obligations. For example, the Philippines’ Fisheries Act forced automatic suspension of extraction and trade for all CITES listed-species, regardless of the Appendix. In contrast, Malaysia suspended trade actively when confronted with a Significant Trade Review for seahorses, apparently concerned about making reliable non-detriment findings. Other countries have suspended trade for other taxa under similar circumstances, as with some stony coral species in Indonesia (UNEP-WCMC 2007). Yet another source of trade suspensions were CITES decisions to suspend exports from CITES member countries that were having difficulty meeting their obligations, as for Hippocampus kuda from Vietnam (2013) and for Hippocampus algiricus for Guinea and Senegal (2016). Eventually Thailand, too, suspended seahorse exports, in January 2016 (too late to be in the database at present), when confronted with ongoing challenges in achieving remedial measures recommended by CITES.    (3) Some trade proceeded without permits, disappearing from formal statistics   63 With respect to the third response, of illegal exports, it is clear that there were regulatory failures in some countries. For example, Project Seahorse trade surveys in Thailand identified considerable illegal and unreported trade at the borders (Kuo et al. unpublished data).   Shifts in the declared trade routes of dried seahorses after CITES implementation highlight the importance of enhancing tracing specimens in trade in order to guide conservation efforts (Nijman and Shepherd 2010, Wu 2016). Our study found increasing proportions of seahorses imported from African countries to Hong Kong SAR and Taiwan, similar to the trend found in sea cucumber trade (Anderson et al. 2011). Given that our gravity model analysis showed that the import countries have strong preferences for sourcing seahorses from regions closer to them, sourcing animals from farther regions might reflect serial exploitation or/and over-exploitation of local resources (Scales et al. 2007, Anderson et al. 2011). It may also reflect established routes for other goods and services. While mainland China, another large market for seahorses, has enhanced its trade relationship with African countries in recent years, it’s possible that mainland China could also expanded its seahorse sourcing from Africa as with other natural resources, such as crude oil and timbers (Bosshard 2008). In addition to the changes in source countries, Hong Kong SAR Customs showed that China was no longer the biggest destination of seahorse re-exports from Hong Kong after 2002. Perhaps traders transporting seahorses to mainland China stopped reporting, or perhaps Hong Kong’s exports were transported to other countries, as seen in shark fin trade (Wu 2016). All the changes in the complex trade routes demonstrated a dynamic trading network that needs continuous and close tracking.    64 Despite the declining recorded trade volume, increasing prices for seahorses suggest further threats to seahorse conservation (TRAFFIC 2008). Hong Kong Customs’ data showed an increasing trend in the seahorse import prices, consistent with several on-site observations of seahorse trade (Kuo et al. unpublished data, Vaidyanathan et al. unpublished data) and trends in prices of other endangered animals (Hall et al. 2008, Challender et al. 2015a). The coincidence of increases in prices with decreases in recorded exports suggest that the demand was not satisfied, and may still be growing with China’s thriving economy (Zhang et al. 2008). While rare wild animals were often popular in the luxury markets, the higher prices on endangered species could provide greater incentives for humans to continue extracting the species (Courchamp et al. 2006). Although seahorses are mainly sourced incidentally, previous studies have indicated that high price could motivate a shift to target fishing, as has happened with trawls in Vietnam (Stocks et al. unpublished data), thus raising conservation concerns (MacMillan and Han 2011).   Our analysis showed that some countries with very limited investment in sustainable fisheries were exporting seahorses, posing questions about implementation of CITES at a national level. We found that seahorse source countries tended to be more involved in demersal fisheries than other range states, as might be expected from previous studies showing that most seahorses were caught incidentally in demersal fisheries (Lawson et al. 2017). Such sourcing means that to ensure the sustainability of seahorse trade, effective management of the demersal fisheries is critical (Lawson et al. 2017). However, as we did not find a country’s efforts in managing sustainable fisheries correlated with changes in its seahorse exports, continuing seahorse trade under CITES regulations might be more determined by economic incentives and trade inertia   65 (Barden 1994). CITES member countries need to pay full attention to meeting the requirement that they make non-detrimental findings for seahorses, as for other taxa (Abensperg-Traun 2009).  Comparisons of different source countries and the results of gravity models among datasets provide insights about potential sources of under-reporting. In contrast to Taiwan’s imports and CITES global trade records, Hong Kong’s imports were inversely related to the number of fishers in the source countries. This is mainly because Hong Kong reported small imports from countries with large numbers of fishers (e.g., Brazil and Indonesia) which did not appear in Taiwan’s Customs and the CITES database. The small reported imports from those countries, comparing to other countries with similar number of fishers, may be explained two ways: (1) imports from these countries are very under-reported /and or (2) these countries retain a significant amount of seahorse catch in their domestic market, such that it is not detected in international trade records. The latter would certainly be likely of Indonesia and perhaps of Vietnam (Vincent 1996). Domestic trade also distorts global trade figures for mainland China. In the CITES database, it reports importing fewer seahorses than much smaller Hong Kong SAR and Taiwan. It turns out that China’s consumption of ten tonnes of seahorses annually is largely derived from domestic waters (Han 2013).  The trade volume documented in Customs records and CITES trade database might be largely underestimated, since the declared trade data did not include (probably increasing) illegal and unreported trade. Although seahorse imports must be reported to Customs, they are easily carried in personal luggage that bypasses inspection. In 2013, at least 700 kg seahorses were seized at Hong Kong airport, suggesting there might be more seahorses have been smuggled in to Hong   66 Kong markets without records (Oriental Daily 2014). In addition, studies on other wildlife have shown that illegal trade could continue even when the exports were suspended, and those trades would not be captured by official data (Challender et al. 2015b). While the CITES trade database only recorded an average of seven million dried seahorses in trade, a meta-analysis shows that at least 37 million seahorses were caught incidentally every year (Lawson et al. 2017). While some of those seahorses might be discarded, most probably went into domestic trade (in just a few countries) or were exported illicitly (Lawson et al. 2017). We should therefore always interpret the trend in documented trade with caution, and be aware that it represents minimum trade. That said, declared trade data, though underestimated and full of discrepancies, certainly has great potential for systemically and quantitatively examining policy impacts on trade on a large temporal and spatial scale (Bruckner 2001, Anderson et al. 2011, Patel et al. 2015).  Having shown that an international environmental agreement affected recorded trade, we will need to deduce whether such agreement reduced pressure on wild populations. A species listing on Appendix II is merely a call to action. The value of the listing lies in its implementation. Thus far most legal trade in seahorses has been suspended, either voluntarily or involuntarily. For example, outside the time frame covered by the reported CITES data, CITES decided to suspend exports from Senegal and Guinea for one seahorse species, in January 2016. This, combined with the closures of exports from Malaysia, the Philippines, Thailand, Vietnam (one species), and other countries creates a situation where 96% of previous global trade in seahorses would not be permitted. The question is whether this trade has indeed ended or merely been redirected through illegal channels. Given the ongoing capture of huge numbers of seahorses in nonselective gear, the former seems more likely, which highlights the need to find ways to address bycatch of   67 species listed on the Appendices. At the moment, specimens sourced this way are not considered in the main CITES enforcement mechanism, the Review of Significant Trade (RST). More effort needs to be applied to analyzing and combating illegal, unregulated, and unreported trade and to addressing bycatch sourcing for marine fishes.   68 Table 4.1 Results of testing the effects of CITES interventions using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂𝟏𝒀𝒆𝒂𝒓𝒕 +𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰). We tested each year from 2003-2007 for Hong Kong’s and 2000-2007 for Taiwan’s data as the break in each model. Here only we show the results from the model with the highest r-square, with p-value of each coefficient estimate in the brackets. The results for other models (with break point as other years tested) were shown in Appendix B. HK: data from Hong Kong CSD (1998-2014); TW: data from Taiwan Customs (1983-2014).             Breakpoint (year) r2 Year, 𝑎1 Slope change, 𝑎2  Intercept change, 𝑎3 Annual weights of imports (kg) HK 2005 0.92 2521 (<0.01) -3807 (<0.01) 7614557 (<0.01) TW 2001 0.78 323 (<0.01) -1017 (<0.01) 2033704 (<0.01) Number of source countries HK 2005 0.92 0.57 (0.05) -0.88 (0.01) 1756 (0.01) TW 2005 0.80 -0.04 (0.30) -0.21 (0.11) 417.33 (0.11) Evenness in supply HK 2006 0.82 0.00 (0.03) -0.00 (0.14) 7.53 (0.14) TW 2002 0.63 0.00 (<0.01) 0.01 (<0.01) -12 (<0.01) Prices (USD/kg) HK 2007 0.61 -0.00 (0.96) 0.25 (0.04) -0.00 (0.04)   69 Table 4.2 Results of gravity model for Hong Kong dried seahorse imports (Hong Kong CSD data), Taiwan’s dried seahorse imports (Taiwan Customs), and global bilateral dried seahorse trade (CITES data). Hong Kong and Taiwan’s imports in pre-CITES implementation (pre 2005) and post-CITES implementation (2005-2014) periods were separated in order to compare to the CITES data (2005-2014). Sample size (n) including trade volume = zero (set as 10-10) for potential country pairs in trade (see Method). All variables are in logarithm space.      Hong Kong  1998-2004 Taiwan 1983-2004        Estimate SE P-value Estimate SE P-value      (Intercept) 9618.40 36386.75 0.79 -6491.53 31318.92 0.84    Source # of fishers  -4.49 1.36 <0.01 1.64 0.69 0.02    GDP per capita  -3.99 0.88 <0.01 -0.43 0.64 0.5    demersal fish catch  3.13 1.32 0.03 -0.56 0.76 0.47    Destination  # of fishers  301.65 382.4 0.43 -49.14 328.13 0.88    GDP per capita  56.37 86.72 0.52 1.8 7.27 0.81    demersal fish catch -17.02 12.65 0.18 8.75 6.96 0.21     Distance -3.47 1.47 0.02 -5.23 0.92 <0.01      Year -1659.54 5142.56 0.75 922.56 4651.58 0.84           n   189     352        70      Hong Kong  1998-2004 Taiwan 1983-2004    R-square   0.2     0.2          Hong Kong 2005-2014 Taiwan 2005-2014 CITES (global) 2005-2014     Estimate SE P-value Estimate SE P-value Estimate SE P-value   (Intercept) 17141.01 56215.85 0.76 6468.4 31395.8 0.84 522.35 154.64 <0.01 Source # of fishers  -4.84 1.03 <0.01 0.16 0.76 0.83 -0.16 0.19 0.38 GDP per capita  -3.39 0.61 <0.01 -0.43 0.72 0.55 -0.49 0.16 0.02 demersal fish catch  4.66 0.93 <0.01 0.91 0.72 0.21 0.62 0.23 <0.01 Destination  # of fishers  290.4 825.11 0.73 -471.88 963.89 0.63 0.06 0.14 0.64 GDP per capita  -26.81 57.18 0.64 17.47 26.79 0.52 0.21 0.32 0.50 demersal fish catch 15.98 33.4 0.63 -3.71 12.44 0.77 0.37 0.15 0.01  Distance -3.52 1.03 <0.01 -0.33 1.07 0.76 -3.14 0.84 <0.01   Year -2582.88 8280.83 0.76 -85.29 5326.28 0.99 -0.26 0.08 <0.01 n   270     160     1639   R-square   0.26     0.07     0.05      71  Figure 4.1 Changes in seahorse trade (in kg) over time. Import data from Hong Kong CSD and Taiwan Customs are compared to the reports from export countries (RE) and import countries (RI) in the CITES database for years after 2005. The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled.  72  Figure 4.2 Number of source countries of dried seahorses for Hong Kong and Taiwan. Data from the CITES trade databases considered reports from both export and import countries (Exports from imports). The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled.   73  Figure 4.3 Eveness (Gini Index) of the supply of dried seahorses for Hong Kong, Taiwan (CSD/Customs data + CITES data) and global trade (CITES data). The higher Gini index indicates less even in the supply. The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled.   74  Figure 4.4 Changes in imports from each source country for (a) Hong Kong and (b) Taiwan. To see the changes in the proportions of imports from minor source countries, we show the cumulative imports to 40% of total imports in (c) and (d), for Hong Kong and Taiwan respectively. The year of CITES listing (2002, CITES1) and implementation (2004, CITES2) are labeled.  75  Figure 4.5 Changes in proportions of dried seahorses imports to Hong Kong from each source country and re-exports to each destination country. (a)-(h) shows the relative imports and re-export for every two years, and (i) shows the absolute import and re-export weight overtime.  76  Figure 4.6 Changes in import prices of dried seahorses for (a) Hong Kong and (b) Taiwan. Prices for seahorses from each source country are presented in different colours. Lines show the mean price of all source countries for each year, with 95% confidence intervals in gray.  77  Figure 4.7 Comparisons of dried seahorse source countries (blue, n=32) and non-source countries (black, n=68) on the space of the first two principle components. Variables used in the PCA included demersal fish catch (Demersal catch), general trade value with China (Trade w. China), per capita GDP (GDP p.c.), marine fisheries employment, and distance to China (Distance).  78  Figure 4.8 Comparisons of seahorse source countries that stop exporting seahorses after the CITES listing (blue) and the countries that continue their exports (black). Variables used in the PCA included demersal fish catch (Dcatch), per capita GDP of each source country (GDP), the average seahorse trade volume in the pre-CITES period (preCITEStrade), marine fisheries employment (Employment), subsides for fisheries capacity building (sub.capacity) and beneficial fisheries subsides (sub. Beneficial).   79 Chapter 5: Investigating patterns and changes in the declared trade of CITES Appendix II animals across taxa  5.1 Synopsis Increasing international trade has posed significant threats to wildlife. The Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) was established in 1975 to address such threats, either by ending or regulating exports. For the more than 4,000 animal species listed on CITES Appendix II, their trade among 182 countries must be limited to levels that are not detrimental to the survival of the species in the wild. We used the official CITES trade database, which includes both export and import data, to evaluate the trade of Appendix II animals among these countries from 1991-2014. We aimed to (1) characterize and compare the trade networks between marine species and other taxa, and (2) identify the major export/import countries at a multi-taxa and global scale by considering the number of species in trade and trade connections among countries. We found that more countries were involved with the trade in marine fishes and invertebrates (except corals) and had more connections than trade in other taxa with a similar number of species. We also found that many countries reported trading in a high number of non-native species, indicating reporting errors in the CITES data. We discovered a South to North trade for all animal species on Appendix II; in which the USA was the centre of reported trade from 1991-2014, and some Asian countries became notable importers. Our study can help prioritize conservation efforts at global scale and guide future trade regulations.    80 5.2 Introduction Poorly regulated international trade in wildlife accelerates the exploitation rate of species, posing significant threats to their survival. The most recent estimate in 2005 suggested that more than 35,000 species are involved in the international wildlife trade, with a global value of more than 300 billions USD (Engler 2008). Wildlife are traded in diverse forms to serve various purposes, including hunting trophies, substrates for artwork and trinket manufacture, bush meat, skins, medicinal products and live as pets (Oldfield 2003, Nijman 2010). However, unsustainable international wildlife trade raises severe conservation concerns. With growing demands alongside globalization, over-exploitation has resulted in many species being classified by IUCN as endangered or extinct (Courchamp et al. 2006, Phelps et al. 2010). The unsatisfied demand may not be thwarted by declining resources and increasing prices (Courchamp et al. 2006). Instead, consumers may turn to alternative resource-supplying countries, often resulting in serial exploitation (Anderson et al. 2011). Controlling cross-border wildlife trade and over-exploitation requires efforts from local and national governments along with international joint ventures.  The level of extraction and supply of wildlife to international markets varies among countries, and is determined by consumer preferences, regional biodiversity, and the capacity of local governments in managing sustainable trade (Abensperg-Traun 2009, Nijman 2010). The international demand for different wildlife products may cause different levels of biodiversity threats in different regions. For example, tropical countries such as Indonesia have suffered great deforestation because of global timber demands (Barbier et al. 1995), while China is the biggest reported producer of seafood from capture fisheries (FAO 2014). Net consumers, mainly developed countries, form a large proportion of the biodiversity footprints linked to developing   81 countries (Lenzen et al. 2012a). The asymmetric market structure of international trade indicates uneven environmental pressures on different regions (Watkins and Fowler 2002, Rice 2007, Moran and Kanemoto 2017). The role of national government therefore becomes vital in protecting local resources. Governments may take various measures to prevent over-exploitation, including quotas and entry limitations to restrict accessibility to their natural resources. However, the political will of governments, capacity for conservation, and governance vary among countries, resulting in different levels of conservation success (Kaufmann et al. 2009, Jones et al. 2013).   The international community has been working together to address the threats on biodiversity driven by international trade. One of the biggest and oldest multilateral environmental agreements, the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), was established in 1975 to ensure sustainable international wildlife trade by ending or constraining exports, and now involves 182 countries as members. CITES manages the species in trade through its three Appendices. While exports in species listed in Appendix I are basically prohibited, exports of Appendix II species must be monitored and controlled to ensure sustainability. Roughly 21,000 of the 35,500 species covered by CITES are listed in Appendix II, which means that their exports must be sourced legally and limited to levels that are compatible with the survival of the species in the wild. If a country fails to prove their exports are sustainable, it may lead to suspensions of the trade. CITES aims to facilitate cross-boundary conservation collaborations, and develop a large-scale, transparent wildlife trade monitoring program.     82 Monitoring wildlife trade can provide rapid information on the changes in supply/demand, thus improving responses to conservation concerns. Market surveys have a long history of being used to monitor wildlife trade (Round 1990). Market surveys provide in situ observations about the market, yielding detailed qualitative and quantitative information (Lee et al. 2005, Nijman and Shepherd 2007, 2011a). In some cases, information from market surveys have been combined with customs data to better quantify the trade and trace changes in volume, value, species or/and participants involved (Chen et al. 2009, Lam et al. 2014). However, these methods either require intensive time, personnel, and investigation, covering only a few taxa, or are constrained to a smaller spatial and temporal scale. The CITES trade database maintained by UNEP-WCMC (The United Nations Environment Programme’s World Conservation Monitoring Centre) provides an alternative and unique source of data to understand and detect changes in wildlife trade. CITES requests all Parties to report their annual trade of the listed species, and hence has one of the largest, most comprehensive, spatiotemporal wildlife trade databases. Using this data, several global patterns in wildlife trade have been captured and can provide insights for conservation (Robinson et al. 2015, Foster et al. 2016).  Most studies using CITES data to monitor wildlife trade have focused on a single taxon or country (e.g., Nijman and Shepherd 2011, Robinson et al. 2015, D’Cruze and Macdonald 2015). These studies overlook the fact that the key players and trade routes may differ at different scales (by species, genus, or across taxa), and a large-scale conservation strategy needs to comprehensively consider multiple species together (Grenyer et al. 2006). Regional studies have suggested that several species serve the same purpose (e.g., used for traditional medicine) and could be exported/imported by the same traders (Lange 1998). Identifying the trade routes with   83 the most species involved can help focus conservation efforts on those routes. This is particularly important as the diversity of species in trade has increased in recent decades, at least for some taxa (e.g., reptiles and amphibians), accompanied by observations of growing demand (Nijman and Shepherd 2011, Robinson et al. 2015). Examining the wildlife trade at a global and multiple species-scale can help set conservation priorities.  CITES data proves to be particularly useful for investigating the network of wildlife trade at a large scale. Though CITES data only includes legal trade, it covers a high proportion of the countries involved in the illegal wildlife trade (40-80%, depending on the type of products) (Patel 2015). This data shows a large capacity for describing countries involved in wildlife trade and the connections among them. Previous studies have used network analyses to identify key participants in the trade and provided insight to conservation strategies based on the network properties they identified (Goss and Cumming 2013, Patel et al. 2015). However, those studies have either focused on few iconic taxa (e.g., rhinos, elephants, and tigers) (Patel et al. 2015), or on a small region (Goss and Cumming 2013). An understanding of wildlife trade networks in general and at the global scale is still lacking.     Marine invertebrates and fishes, two of the largest groups of wildlife in trade, have only been included in CITES in recent years, highlighting the demand for more understanding on the trade in marine species (Vincent et al. 2013). Products from capture fisheries compose 25% of the import value of global wildlife trade (Engler 2008). However, with the exception of some invertebrates (e.g., corals and clams), most commercial marine species were thought to be under the scope of FAO and/or regional fisheries management organizations (RFMOs), rather than   84 CITES (Vincent et al. 2013). With increasing threats from over-fishing, CITES has gradually been acknowledged as a complementary tool to protect marine species – since 2002, commercial marine fishes have been listed on CITES Appendix II at every Conference of the Parties. While CITES will be increasingly important for managing the trade in marine species, we need to examine if our experiences with other CITES-listed animals can be applied to the newly listed marine species. Studies on the ecological footprints of international manufactory supply chains have found different hotspots of anthropogenic threats for biodiversity in land and marine ecosystems (Moran and Kanemoto 2017). We therefore hypothesized that the international trade in terrestrial and marine species may also have different geographic patterns, requiring different approaches to conservation.   To understand the international wildlife trade from a multi-species perspective, and probe for patterns in the trading network across taxa, we used the CITES trade database (drawing on years 1991-2014) to evaluate movements of Appendix II species among countries. We were particularly interested in the trade of animals that required CITES export permits with non-detriment findings (an assessment results by a Scientific Authority that the export of the species will not be detrimental to the survival of the wild populations). Those specimens were either sourced from the wild or born in captivity from wild parents. We aimed to (1) characterize and compare the trade networks between marine species and other taxa, and (2) identify the major export/import countries at a multi-taxa and global scale by considering the number of species in trade and trade connections among countries. For the second aim, we first determined the major range states (countries where a species is known to live), exporters (countries that trade a species with other countries), and importers for CITES Appendix II listed animals. Then, we explored   85 changes in relative species richness in exports/imports over time for each country. Finally, we examined changes in countries’ connections over the time period considered for the trade in all CITES Appendix II animals. We discuss our findings in light of current understandings in international wildlife trade and provide recommendations to support the prioritization of conservation efforts for sustainable trade.  5.3 Methods 5.3.1 Data We retrieved trade records for all animals listed on CITES Appendix II, reported by exporter, importer, or both, from the CITES trade database (http://trade.cites.org, accessed on February 9th, 2017). We accessed trade records for all species in Kingdom Animalia listed in the CITES Species Checklist (http://checklist.cites.org/, accessed on July 26th, 2016). Among all the reported transactions, we focused on trade for commercial use, labelled as “T” for purpose in the CITES trade database (UNEP-WCMC 2013b). To avoid double counting, we excluded transactions that were reported as re-exports (had not been sourced in the reported exporting country). We focused on trade that required non-detriment findings (i.e., directly taken from the wild, ranched wild animals, or first generation captive born individuals with wild-caught parents) since such trade is most likely to directly affect wild populations. Since information on the source of animals in trade was seldom recorded before 1991 (UNEP-WCMC 2013b), our study only used data from 1991-2014, extracting a total of 186,824 entries.   In our study, we included both imports and exports data, and treated all trade records equally (whether they were reported by an exporter, importer, or both). Our study was probing the   86 general patterns of trade in terms of species, countries involved, and connections among trading partners. We set aside trade quantities in this cross-taxon study, since they were too difficult to standardise across all the different taxa and units found in CITES records (including kg, pieces, boxes, individuals, etc.). In contrast, we focused on the other reported trade information, such as connections between countries and species in the trade. All countries/territories reported in the CITES database are under the official ISO-alpha-2 codes (https://www.iso.org/obp/ui/#search), and “XX” denoted an unknown exporter or importer. We have attached a table showing the ISO codes and countries’ full names for those countries mentioned in this paper (Table D1).  5.3.2 Analysis Our study explored the following aspects of the trade of CITES Appendix II animals between 1991 and 2014: (1) comparison of the basic properties of the trade network among taxa; (2) identity of key range state, exporting and importing countries across taxa; (3) identify the discrepancies between range-states (countries where a species is known to live) and exporting countries for any given species; (4) changes in the number of species reported in trade; and (5) changes in the countries’ trade connections.   In analysis (1), we examined the trade network across all CITES Appendix II animals and also for seven major groups of taxa separately, for cross-taxa comparisons. The seven groups of taxa we examined were: (a) Corals (Class: Anthozoa & Hydrozoa); (b) Marine fishes and non-coral invertebrates; (c) Mammals (Class: Mammalia); (d) Birds (Class: Aves); (e) Reptiles (Class: Reptilia); (f) Amphibians (Class: Amphibia); and (g) Insects (Class: Insecta). For (b), we list the species that we considered as “marine” in Table D2. Since there were only two freshwater fish   87 species (Polyodon spathula and Arapaima gigas) reported to CITES in 1991-2014, we did not include freshwater fishes in the cross-taxa comparison.   The analyses (2)-(5) considered the trade for all CITES Appendix II animals. Since analyses (2)-(4) looked at number of species, we only used records that reported animals in trade to species level (156,553 of the 186,824 entries, covering a total of 2,055 species). Only half of the approximately 5,000 animal species listed in CITES Appendix II were reported in trade for the years analysed, and 19% of these were only reported in captive trade and thus were excluded from analyses.   5.3.2.1 Examine properties of the wildlife trade network and compare the networks among taxa We examined the basic properties of the trade network for all animals then for each of the major taxa. We considered the network in two time frames: 1) with all the countries that have ever been reported in the CITES trade database as an exporter or importer of the listed animals from 1991-2014, and 2) with the trade reported in 2014 only. The first time frame aims to capture the most completed network. The second time frame accounts for the biases due to the different years that species were listed, and represents the most recent trade with implications for conservation actions taken onwards.   For each network, we calculated the number of countries in the trade (nodes) and the number of links between countries (degree). We then calculated the reciprocity and density of each trade network, using the “igraph” package in R statistical software (Csárdi and Nepusz 2006).   88 “Reciprocity” measures the proportion of mutual connections in a network (i.e., two countries were exchanging species). “Density” measures the proportion of the number of links over the number of possible links, representing how connected a network is. We were also testing whether the distribution of degree follows a power-law (scale-free network), which was commonly seen in other real-world networks (Watts and Strogatz 1998, Goss and Cumming 2013). A scale-free network, which has only few highly connected countries, is more vulnerable if the most connected country stops exporting or importing (Watts and Strogatz 1998, Goss and Cumming 2013). To test whether the CITES trade network is a scale-free network, we fitted a power-law function to the degree distribution.   5.3.2.2 Key range states, exporting and importing countries For each country, we calculated the number of CITES Appendix II animal species occurred in their jurisdiction. Information about the range state of each CITES species (where the species live) was downloaded from the CITES Species Checklist database (http://checklist.cites.org/, accessed on July 26th, 2016). Then, for each country, we calculated the number of CITES Appendix II animal species it reportedly exported and imported. We considered a country to be an exporter or importer of a species as long as it reported exporting/importing the species at least once between 1991 and 2014.  5.3.2.3 Exploration of the discrepancies between range-states and exporters After mapping the range state countries and exporting countries for all the CITES Appendix II animal species, we examined matches in the country sets. We were interested in two situations: (1) a country that was a range state of a species, but was not reported to be exporting that   89 species; and (2) a country that was not a range-state of a species, but was reported to be exporting the species. Therefore for each country, we calculated (1) the proportion of native species reported being exported, and (2) the number of non-native species reported exported.   5.3.2.4 Exploration of the changes in the number of species a country exported/imported We identified the countries exported/imported the highest number of species in 2014, to understand the most recent trade and target the countries that need conservation focus. To do so, we tallied the number of species each country reported exporting and/or importing in 2014, and identified the ten largest exporters/importers. Then, to investigate how the species richness in trade changed over time, for each of those 10 countries, we created an index of the relative number of species exported/imported by the country for each year from 1991-2014. This was calculated by comparing the number of species in its exports and/or imports with the total number of species in global trade - across all countries, not just the top 10 - for each year. This was necessary to adjust for the changing number of species listed on Appendix II over time. We then applied linear regression on the relative species index versus year for each country to examine the trend of species richness in trade.  5.3.2.5 Exploration of the changes in countries’ connections for the trade in CITES Appendix II animals  We calculated three indices to measure a country’s connectivity with other countries from 1991-2014 for all CITES Appendix II animals: (1) out-degree (2) in-degree, and (3) closeness centrality (Freeman 1978). Higher out-degree, in-degree and centrality indicate a country has a higher connectivity with other countries.    90  For out-degree, we tallied the number of countries to which a particular country exported, for each year. For in-degree, we tallied the number of countries from which a particular country imported for each year. For closeness centrality, we calculated the mean number of links it takes to connect any one country to other countries trading any animal species listed on Appendix II. This was done between all possible pairs of trading countries, treating imports and exports as equal trade links. Two countries with direct trade were connected by one link. For trade between two countries that had no direct link, we discerned the shortest path through the network of countries between these two, and calculated the number of links along the path. For any two countries that were not connected through a path, we assigned the distance between the two the maximum possible distance in the network plus one. We then calculated the mean trade links for all pairings for a particular country, thus deriving an index of connectedness for the country. We considered those countries with smallest mean number of links to be the highest ranked in terms of closeness centrality.  To see the changes overtime, we repeated the calculations for all years from 1991-2014. We then ranked all the countries based on the three indices, separately, and identified the top 10 for 2014. Again, we focused on the 2014 was for the most recent trade with all the species that were listed in different years. Next we examined the trend in the connectivity for these 10 countries, for each index, by testing the significance of the slope of out-degree/in-degree/centrality versus year using linear regression.    91 5.4 Results 5.4.1 Examine properties of the wildlife trade network and compare the networks among taxa From 1991-2014, the trade in CITES Appendix II animals involved 238 Parties with 4,015 links among Parties (Table 5.1a). Note that the number of Parties involved in trade is higher than the number of CITES Parties, because of reports on trade with non-CITES Parties (e.g., Saint Kitts and Nevis), dependent territories (e.g., Faroe Islands), and former countries (e.g., Yugoslavia). Among all taxa, the trade in reptiles had the highest number of countries involved and had the highest number of links (Table 5.1a). For marine fishes and non-coral invertebrates, we found that despite fewer species in international trade (n=38), more countries participated in the trade than other groups of taxa with similar number of species in trade (amphibians and insects) (Table 5.1a). In addition, the trade networks of taxa groups with fewer species in trade (marine fishes and non-coral invertebrates, amphibians, and insects) have lower reciprocity (Table 5.1a). We also found that the trade network of corals had the lowest density, indicating the lowest connectivity (Table 5.1a). The trade in CITES animals, either across taxa or by taxa, were scale-free networks, meaning few countries were much more connected than other countries (Table 5.1a).   We found that the trade networks in 2014 have similar patterns among taxa as the ones across years (Table 5.1b & Figure 5.1). The trade network of 25 marine fish and non-coral invertebrate species had a total of 148 links, while the 102 bird species in trade had only 131 links (Table 5.1b & Figure 5.1). In contrast to the trade networks across years, the reciprocity rate and density   92 in amphibian trade was higher than other taxa with similar number of species in trade (Table 5.1b). All the trade networks in 2014 were scale-free (Table 5.1b).     5.4.2 Key range state, exporting and importing We found that Indonesia (ID) was the range state for the greatest number of CITES Appendix II animal species (n=1,367), compared to other countries. It was followed by Australia (AU, n=969), the Philippines (PH, n=845), Japan (JP, n=782), and Papua New Guinea (PG, n=776) (Figure 5.2). Indonesia was also reported as exporting the highest number of animal species in 1991-2014 (n=640), followed by the United States (US, n=630), Australia (n=356), Malaysia (MY, n=216), and Fiji (FJ, n=201) (Figure 5.3a). For the same time period, the countries reported importing the highest number of Appendix II animal species were mainly developed countries: the United States (n = 1,496), Japan (n=1,051), Germany (DE, n=858), the United Kingdom (GB, n=752), and France (FR, n=685) (Figure 5.3b).   5.4.3 Exploration of the discrepancies between range-states and exporters We found discrepancies between range-states and exporters. Although range states with high CITES species richness generally exported more species, the correlation coefficient between the two was only 0.6. Most of the range-states that exported a high proportion of local species were developing countries. Among the top 20, Guinea (GN) exported more than half of its listed species (62%), followed by Togo (TG, 44%), Cameroon (CM, 40%), Denmark (DK, 38%), and Indonesia (37%) (Figure 5.4). A number of countries reported exporting Appendix II-listed species that were not native to them (Figure 5.5). The United States exported 536 non-native   93 species, followed by Canada (CA, n=155), Switzerland (CH, n=135), Indonesia (n=131), and South Africa (ZA, n=112) (Figure 5.5).   5.4.4 Exploration of the changes in the number of species a country exported/imported Indonesia, the United States and Australia exported relatively higher number of listed species than other countries in the same year, together accounting for a mean of 50% of overall trade across 1991-2014 (26%, 16%, and 7% respectively). Most of the top 10 countries had exported more species over time (Table D3 & Figure 5.6a). Among them, we found Australia to have the highest increase in the number of species exported from 1991-2014, and becoming the second most important exporter behind Indonesia in 2014 (Table D3 & Figure 5.6a). The relative number of species in Suriname’s (SR) and Madagascar’s (MG) exports decreased over time, while the relative number of species in Guyana (GY) and Canada’s exports had no significant changes (Table D3 & Figure 5.6a).   For imports, the United States, Japan, and the Germany imported relatively higher numbers of species than other countries from 1991-2014, with a mean of 61%, 47%, and 33% of total species in the global trade. Six of the top 10 import countries imported more species over time, while the Netherlands (NL) reported importing fewer species, and Japan, the United Kingdom, and France had no significant changes in the number of species in their imports (Table D4 & Figure 5.6b). Hong Kong (HK) and South Korea (KR) had the highest increasing rate in the relative number of imported species, becoming the third and fifth ranked importer by 2014 (Table D4 & Figure 5.6b).    94 5.4.5 Exploration of the changes in countries’ connections for the trade in CITES Appendix II animals  For the trade in all CITES animals, the United States and Indonesia had the highest out-degree, suggesting they exported to the greatest number of countries, in both 2014 as well as across 1991-2014 (with a mean of 62.8 and 49.8, respectively) (Figure 5.7a). The 10 countries with the highest out-degree exported to increasing numbers of countries over time, except for Tanzania (TZ), whose exports stayed relatively stable between 1991 and 2014 (Table D5 & Figure 5.7a).   The United States and Japan had the highest in-degree in 2014 and across 1991-2014 (mean = 75.1 and 45.5, respectively) (Figure 5.7b). The in-degree of most of the top10 countries increased over time, although the United States, France, and the United Kingdom showed no significant trend (Table D6 & Figure 5.7b). Among all, mainland China (CN), Hong Kong, and South Korea had the highest increasing rate in in-degree (Table D6).  For centrality, we found that the United States was always the “most central” country from 1991-2014 for the trade in all CITES animals (Figure 5.8). The centrality ranking for Indonesia, South Africa, and Hong Kong increased from 1991 to 2014 (Table D7 & Figure 5.8).   5.5 Discussion  Our analyses identified the distinctive characteristics in the trade network for marine fishes and non-coral invertebrates, and countries that played a pivotal role in trading animal species listed in CITES Appendix II. By identifying important countries/regions in CITES trade, our study may help CITES Secretariats to prioritise future financial and personnel supports for trade   95 management and conservation. We also found that key countries participating in wildlife trade were not necessarily the range states with high CITES species richness. Our finding that developed countries were greater consumers of wildlife products than developing countries is consistent with other studies on the trade in natural resources, suggesting that the responsibility for wildlife protection should be shared accordingly (Rice 2007, Swartz et al. 2010). In addition, our study supports calls for special attention being put on emerging Asian import countries (Nijman 2010, Challender et al. 2015a) and on the trade in marine species (Swartz et al. 2010, Vincent et al. 2013). Despite imperfections in CITES data (Blundell and Mascia 2005), patterns emerging from our analyses across species, countries and time highlight the value of monitoring trade records in conservation decision-making (Blundell and Mascia 2005, Foster et al. 2016). The main caveat is that our work deals with patterns in trade flows and not with specific volumes in trade for one or more taxa, which might demonstrate quite different patterns.  Our study identified distinctive trade network and disparate set of exporters for marine fishes and non-coral invertebrates, implying the need of special management concerns. Compared to other taxa with similar number of species in trade, the network of the trade in these CITES marine species, excluding corals, involved more countries and was highly connected. One of the conservation implications of this finding is that if we expect import countries to provide incentives to, or put pressure on source countries to export sustainably, it must be done broadly. On one hand, actions by importers need to be coherent, collaborative and harmonized to be effective, as source countries have many alternative destinations. On the other hand, more export countries require greater efforts in monitoring and manage the supply of marine species. Our study also suggests that the major exporters of marine species were different to the exporters of   96 non-marine animals, requiring alternate or complementary targeted trade management and conservation actions. For example, we found that Australia and Oceania countries exported many CITES marine species to many countries, while the major exporters of other taxa were Asian, African, or/and South American countries. However, the major exporters of marine species identified in this study were also different to the big export countries of marine capture fishery products reported by FAO and Sea Around US. This is not surprising because most commercial marine fishes are not currently covered by CITES (FAO 2016). To have a broader picture for the trade in marine species, it is necessary to combine data sources from CITES, FAO/Sea Around Us, and regional fisheries management organizations (RFMOs).  In contrast to many studies focusing on single species or regions, our findings cut across taxa to generate broad understandings of wildlife trade at the global scale (Blundell and Mascia 2005, Nijman et al. 2011, Foster et al. 2016). We found that the range states for a high number of Appendix II animal species were mostly the mega-diversity countries, which also were identified as having a high number of endemic species (Mittermeier et al. 1997, Brooks et al. 2006). Some countries with high CITES species richness that were not mega-diverse, e.g., Mozambique, Thailand, and Japan, also include biodiversity hotspots in their jurisdictions (Brooks et al. 2006). Orme et al. (2005) and Grenyer et al. (2006) suggest that the hotspots of total species richness, threatened species richness and endemic species richness are not congruent, while our findings show concordance in the distribution of CITES species and other levels of species richness. This discrepancy might be a result of differences in study scale. First, while the two previous studies examined species richness at area-grid level, our study was conducted at country-level for political relevance. In addition, those studies focused on birds, mammals, and amphibians   97 separately, while we examined the key range states for CITES species across taxa. Our work also focused on species for which exports pose a potential threat while both other studies used IUCN Red List assessments, which include diverse threats. Moreover, the leading threat for many marine species (which we included) is exploitation, while habitat loss is the primary threat for many terrestrial species (Gibbons et al. 2000, Brooks et al. 2002), the focus of previous surveys. Our study provides insight into prioritizing management efforts on regions that can cover the highest number of CITES-listed species, supported by research in favour of focusing conservation efforts using multi-taxonomic approaches (Kremen et al. 2008).  Our research highlights discrepancies between range states and exporters of CITES Appendix II animals, which raises some concerns. We found that the number of CITES species exported by a country was not proportional to the species richness in the range state, which might be due to the different abundance of the species in each country, the costs of extracting/exporting the species, and the ability countries have in managing their trade (Carpenter et al. 2005). Given that CITES requires all listed species’ exports to be accompanied by Non-Detriment Findings (CITES 1973), one might expect that countries with better management of wildlife trade to be exporting more CITES species. However, our analyses showed an opposite trend. We found that most countries exporting a high proportion of native CITES species actually seem to lack capacity to manage their trade adequately. CITES had issued trade suspensions, as a result of inadequate compliance with the Convention, for six of the 10 countries exporting the highest proportion of native CITES animal species in 2014 (CITES Secretariat 2017, accessed at March 20th, 2017). Most notably, CITES had suspended all trade for Guinea, the country reported to be exporting the highest proportion of its native CITES-listed species (CITES Secretariat 2013). This suggests that   98 CITES needs to be more rigorous in examining current permitted trade and countries’ non-detrimental findings. In addition, some countries, especially the United States, reported exporting a high number of non-native species. Approximately half of those species were reportedly only F1 individuals, raised in captivity from wild-caught parents, but the rest were presumably re-exports without proper documentation. The first case raises risks of invasive species unless trade transactions are very carefully controlled, while the second raises questions about the reliability of record keeping (Smith et al. 2009, Herrel and Meijden 2014).   We found that the patterns of international wildlife trade exerted uneven environmental pressure on developing and developed countries, similar to other international trade (Giljum and Eisenmenger 2004, Rice 2007, Swartz et al. 2010). For example, the countries exporting the highest number of CITES species were mainly developing countries, while the main importers were developed countries. This pattern is consistent with observations in the trade of seafood products and some cash crops (e.g., coffee and palm oils) (Chichilnisky 1993, Swartz et al. 2010), as well as the international environmental footprints from exchanging manufactured goods (Lenzen et al. 2012a, Moran and Kanemoto 2017). This pattern of trade flows reflects differences between net exporters and importers (Rice 2007). Possible explanations lie in developing countries’ greater biodiversity, shorter history of resource depletion, greater need of export revenue, greater difficulty in regulating exports (Chichilnisky 1993, Giljum and Eisenmenger 2004). The corollary is that developed countries often have lower biodiversity, have longer histories of large-scale depletion of biodiversity, have more financial stability, and/or regulate use of their natural resources more carefully (Chichilnisky 1993, Giljum and Eisenmenger 2004). This asymmetric pattern in global wildlife trade suggests that the   99 responsibilities in maintaining sustainable trade should be shared along the supply chain including controlling the consumer demand, instead of only pressuring the source countries (Moran and Kanemoto 2017).   While the United States has been the centre of the reported wildlife trade through time, our research suggests that we should focus conservation efforts on emerging trade leaders in Asia. The United States has reported exporting and importing the highest number of species, and had the greatest number of connections with other countries in trade. This prominent documented trade of the United States, while needing careful attention, may arise from the United States’ consistent and careful reporting. Moreover, the United States have been featured as a key importer of wildlife for several years, to a large extent has had the time and capacity to put in place a number of measures attempting to curb, monitor and/or regulate trade adequately (including effective reporting). The United States reported around 86% of the entries in the CITES trade database mentioning the USA in some capacity. In contrast, another large exporter, Indonesia, reported only 76% of all records citing Indonesia and the second largest importer, Japan reported only 71% of its trade entries. Several Asian consumers have become more important in the wildlife trade. Species richness in imports to Hong Kong, South Korea, and Mainland China have rapidly increased since 2000, while Hong Kong and Mainland China have also become more central to the trade in CITES species. Since booming economics in Southeast Asia and China have been observed correlating with increasing demand of natural resources locally and from abroad (Li and Li 1998, Nijman 2010), precautionary conservation efforts should be applied in those emerging consumers.    100 Our findings, as with other studies based on the CITES trade database (Blundell and Mascia 2005, Nijman and Shepherd 2011, Foster et al. 2016), should be interpreted with caution because it did not include illegal, unreported, and unregulated trade. Comparisons of CITES trade data with Customs records and market surveys have identified inconsistencies in reported trade volumes, participating countries, and species involved (e.g., Blundell and Mascia 2005, Nijman and Shepherd 2011, Foster et al. 2016). Illegal or unreported wildlife trade, which could be more than ten times higher than the reported trade, are of course not captured by CITES data (Challender et al. 2015b). Internal trade, which can be important and contribute significantly to declines in species are also not included in the CITES database (Milner-Gulland and Bennett 2003). In addition, the data reported to CITES might represent the trade permitted by national authorities, but not the real volume in trade (UNEP-WCMC 2013b). Despite these sources of uncertainty and under-estimates, CITES data may still be the most comprehensive dataset for international wildlife trade and analyses based on these data provide meaningful insights into finding effective solutions for sustainable trade.  Our study focused on species richness and connections among countries but not trade volume, which might result in different patterns. Using different metrics (e.g., number of species, connections, trade volume) could lead to different results when determining the importance of a country in trade, since those metrics may not correlate with each other. For example, reptiles comprised 29% of the species in the international wildlife trade of Western Cape, South Africa, but only 4% of its total trade volume (Goss and Cumming 2013). However, CITES trade database does not require standardized unit in reporting, thus many reports were in the units that were hard to compare to other reports (e.g., box, bags, bottles). Furthermore, the same number of   101 individuals extracted for trade may imply different levels of conservation concerns to species with different life history traits (Hutchings et al. 2012). In a cross-taxa analysis, it would be difficult to interpret the conservation meaning of the accumulative trade volume across taxa.    Our approaches for analyzing CITES Appendix II species could be applied to trade in non-CITES species of conservation concern. Their survival may be threatened globally or locally, perhaps even partly by trade, and they may be protected by local governments or other environmental agreements (Nash 1993, Anderson et al. 2011, Auliya et al. 2016). However, trade data for most such non-CITES species are rarely collected or are very general (e.g., not classified to species level) (Schlaepfer et al. 2005, Smith et al. 2009). The CITES data infrastructure may present national governments and regional treaties with a useful framework for recording trade data for other species, and keeping track of the key participants and species as our study demonstrated. National government’s efforts in monitoring the trade of some species have benefited further conservation actions (Eriksson and Clarke 2015), and should be expanded to more species in trade.   Monitoring and controlling wildlife trade is critical to maintaining wildlife populations, and often preferable (even in conservation terms) to banning exploitation and use. The trade in wildlife often represents an important source of livelihood to people with few alternatives and may be integral to community culture (Roe et al. 2002, Cooney and Abensperg-Traun 2013). Prohibiting the use and trade of wildlife will commonly not stop exploitation, and instead may well promote illegal trade (Rivalan et al. 2007) making it harder to monitor. Increasing the transparency and reliability of data pertaining to the wildlife trade is imperative for policy   102 makers to make informed conservation and management decisions. It is only possible to control the extraction of natural resources at a sustainable level, when management decisions are made based on understanding the magnitude, participants, scale, and trends in wildlife trade.                        103 Table 5.1 The basic properties of the trade network for all CITES Appendix II animals and seven major groups of taxa, for data (a) from 1991-2014 and (b) 2014 only. (a) 1991-2014 Network Number of species in trade Number of countries Number of links Reciprocity Density Degree distribution fits power-law Alpha R2 All App.II animals 2,055 238 4,015 0.25 0.07 0.56 0.60 Corals 671 150 712 0.10 0.03 0.81 0.75 Marine fishes & non-coral invertebrates 38 99 456 0.04 0.05 0.80 0.80 Mammals 227 192 1,307 0.19 0.04 0.76 0.75 Birds 613 193 1,504 0.14 0.04 0.80 0.79 Reptiles 406 207 2,062 0.20 0.05 0.68 0.72 Amphibians 54 55 136 0.06 0.05 0.83 0.73 Insects 42 62 173 0.03 0.04 0.79 0.76   104                 (b) 2014 Network Number of species in trade Number of countries Number of links Reciprocity Density Degree distribution fits power-law Alpha R2 All App.II animals 678 173 1009 0.12 0.03 0.84 0.85 Corals 286 90 264 0.02 0.03 0.90 0.86 Marine fishes & non-coral invertebrates 25 63 151 0.01 0.04 0.93 0.86 Mammals 67 98 235 0.08 0.02 0.98 0.84 Birds 102 72 131 0.03 0.03 1.18 0.94 Reptiles 159 129 492 0.09 0.03 0.85 0.70 Amphibians 31 20 30 0.07 0.08 0.91 0.82 Insects 27 28 35 0 0.05 0.92 0.86   105  Figure 5.1 Trade networks in 2014 for (a) all CITES Appendix II animals, (b) corals, (c) non-coral marine invertebrates and fishes, (d) mammals, (e) birds, (f) reptiles, (g) amphibians, and (h) insects. For (a), only countries with degree higher than average (n=51) were labeled. Line width represents the number of species in trade (in log scale), and line color corresponds to the region of the   106 source country: Asia: red; Western & Southern Africa: brown; Northern & Eastern Africa: orange; Australia and Oceania: blue; America: purple; Europe: green.    107  Figure 5.2 The number of CITES Appendix II animal species inhabiting each country’s terrestrial jurisdictions and/or EEZ.   108  Figure 5.3 The number of CITES Appendix II animal species reported being (a) exported from or (b) imported to each country from 1991-2014.  109  Figure 5.4 The top 20 range states that exported the highest percentage of CITES Appendix II animal species that are native to them from 1991-2014. Countries are coded in official two-letter country codes (ISO2). The same information is displayed in map format in the inset.  110  Figure 5.5 The top 20 countries exporting species that are not native to them from 1991-2014. Countries are coded in official two-letter country codes (ISO2). The same information is displayed in map format in the inset.  111  Figure 5.6 The percentage of all species (a) exported and (b) imported by a country over the total number of species in trade in each year, from 1991-2014. We show the top 10 countries in terms of exporting/importing the highest number of species in 2014, in descending order, and trace how the relative species diversity in trade for these 10 countries changes over time. Countries are coded in official two-letter country codes (ISO2).  112  Figure 5.7 Changes in (a) out-degree and (b) in-degree for the 10 countries found to have the highest out-degree/in-degree in 2014. Countries are listed in descending order of importance along the y-axis.  113  Figure 5.8 Changes in the ranking of centrality from 1991 to 2014 for the 10 countries ranked as having the highest centrality in the CITES Appendix II animal trade in 2014. Countries are ranked along the y-axis in descending order.    114 Chapter 6: Conclusion 6.1 Overview My thesis explored the new frontier of marine conservation where marine fishes that are commercially important are also regarded as wildlife, to the extent that their international trade is regulated for sustainability. My thesis is the first study to examine the effects of this new governance model through multiple approaches, ranging from analysis of fishers knowledge to exploration of international trade flows. Using CITES as a case study, my research advances our current understandings of trade regulations on wildlife to support adaptive management for marine species in trade.   In my first two research chapters (Chapters 2 & 3), I used seahorses as a case study to evaluate national level fisheries and trade implementation of CITES Appendix II for a genus of marine fishes. My findings provided an important snapshot for seahorse trade in a country that was (i) until last year the biggest source for international markets and (ii) under official scrutiny as it sought to implement CITES requirements. In Chapter 4, I broadened my scope to the global level, systematically examining the impacts of CITES on the international trade of seahorses, as the first marine fishes for which CITES regulated trade. The findings of this chapter highlight the importance of monitoring the trade for rapid responses to the changes in trade for effective management. In Chapter 5, I identified the key countries involved in trade of Appendix II-listed species, with focus on marine fishes. My results reveal the importance of particular countries and should help prioritize conservation efforts across taxa. The findings of my thesis confirmed that the management for sustainable use of marine fishes could benefit from the instruments   115 conventionally used for other wildlife, but I also discerned that some unique challenges for marine fishes (e.g., bycatch) will need special concerns.   6.2 Research Findings My thesis shows that the management of the trade in marine fishes could be put into the context of management of any other wildlife trade (Vincent et al. 2013). I used CITES as a case study to investigate the fisheries and trade of marine fishes under export regulations. I approach this objective at multiple scales, from local to international, and from a single genus to all animals listed on CITES Appendix II. In the following sections, I will review how each of my research chapters respond to my research questions and the implications of my research.  6.2.1 Research Question 1: How to get the best quantitative information for data-limited species in fisheries and trade? In showing that estimates for annual catch rates differed depending on the reporting period, this chapter should lead us to improve narrative reports of catches by choosing time periods judiciously. Although fishers’ knowledge can provide critical baseline information for the management of data-limited species (Hagan et al. 2007), its reliability is often questioned (Daw 2008). Some known sources of bias in fishers’ reports are hard to avoid or correct, such as fishers’ tendency to exaggerate their recreational catches (Sullivan 2003) or to under-report for species that are conservation targets (Rist et al. 2010). The hitherto unexplored bias we outline here can, however, be addressed. The best approach will be to undertake pilot studies that compare fisher’s reports for different time periods with known extraction rates. The known rates can be obtained through port sampling or on-board observations (Daw et al. 2011) and fishers’   116 estimates for all time periods can be improved by asking and recording zero events during the interviews (Golden et al. 2013).   The results of my study have implications for all resource assessments based on informant recall. The management of recreational fisheries and hunting, for example, often requires fishers and hunters to report their extraction data. Our findings of the recall bias in the reported catch rates add to the evidence that variability in time periods could help explain inconsistencies in reported fishing days, catch rates, and consumption in recreational and artisanal fisheries, and local consumption rates of bushmeat (Connelly and Brown 1995, Daw et al. 2011, Golden et al. 2013). Further, our findings also raise concerns about data collection for commercial fisheries, where fishers may fill in their catch to logbooks post hoc (Gavin et al. 2010). Other research fields that collect quantitative data by respondents’ recalls, such as public health, medicine, and forensic science, could also benefit from our findings by taking into account the effects of reporting time periods.  6.2.2 Research Question 2: Has CITES impacted the trade in marine fishes at the national level? My research indicates that the accumulated value of incidentally caught fishes can be substantial, providing incentives for the fisheries to continue even once the target species are over-exploited and producing poor returns. Economic theory has predicted that the fishing costs will increase with the decreases in fish population, and the fishing effort will decline once the fisheries is economically not profitable (Gordon 1954). However, my study shows that the economic values of bycatch species can be enormous accumulatively, suggesting that the revenue gained from   117 such incidental catch can help subside the fisheries to continue. My findings add to the emerging literatures on “opportunistic exploitation”, which suggests that the extraction of less desirable species may lead to extinction of the most valuable but (increasingly) rare species (Lobo et al. 2010, Branch et al. 2013). The destruction power of some non-selective fishing gears (e.g., bottom trawl) makes the situation more worrisome, as the continuing fishing not only empties up the oceans but also damages the habitats.   My research reveals the concerns that export regulations per se have limited to no effects on the extraction of bycatch species, and may aggravate difficulties in management by generating an increase in illegal trade (Tolotti et al. 2015). I found that a drop in declared export volume was not accompanied by a decline in the catch of seahorses. This suggests that Thailand should recognize that managing export is not a substitute for managing fisheries, and there is urgent need to realize its conservation commitment to concrete fisheries reforms. Delays in realizing, or limitations on, the effects of international agreements at the extraction level have also been found for other marine fishes (e.g., sharks), suggesting that national governments need considerable encouragement to implement international agreements (Davidson et al. 2016). We are reminded that environmental treaties are a call for action rather than a solution, and different measures should complement each other to achieve the maximum effectiveness (Young 2011).   6.2.3 Research Question 3: Has CITES impacted the trade in marine fishes at a global level? My study shows that documented international trade in seahorses changed after their CITES listing, such that trade monitoring is clearly important for adaptive management. I found that   118 import prices of seahorses rose with declines in declared trade volume, consistent with findings in other wildlife trade studies (Challender et al. 2015a). Increasing prices with rarity of some wildlife can provide incentives for illegal catch/hunting (Courchamp et al. 2006, McClenachan et al. 2016). However, direct linkage between prices and demand for seahorses has not yet been assessed. I also found that a major seahorse importer, Hong Kong, has expanded its sources to include a number of new African and South American countries over time. Although the timing of such changes did not directly coincide with CITES interventions, the expansion highlights the need to focus conservation efforts on new source countries, especially given that many countries that were hitherto major sources for seahorses have officially suspended or banned their exports. This study also found that even under CITES regulations, exports of seahorses were mainly affected by the economic status of the source countries rather than investments in management efforts therein.   My finding that low-income countries tended to export more seahorses is different to the trend found in seafood trade, for which the high-income countries also exported a substantial amount of capture seafood (Watson et al. 2017). That says, exports of seahorses are important for the people in the low-income countries, indicating the need to achieve the sustainable trade for these species (Abensperg-Traun 2009). To do so, enhancing national capacity and willingness to improve management for exploitation and trade of at-risk species is clearly important, and not yet achieved for seahorses. Based on my findings that many CITES member countries continued exporting seahorses without much investment in sustainable fisheries management, their implementation of sustainability provisions of the Convention is open to question. In addition,   119 even for the countries that did suspend their exports, action is still needed to achieve the Convention’s goal of trade without detriment.   6.2.4 Research Question 4: Do marine species have different trade patterns compared to other CITES Appendix II animals? My findings show that more countries are involved in the trade of marine species than for other taxa, indicating the urgent need for international collaboration. In fact, the trade in marine species might be even more complex than CITES data captured; our understanding of trade in the many commercial marine species (including those not included in CITES) is obscured by exchange on the seas, landing at foreigner ports, and processing in other countries before re-export to the final destinations (Pramod et al. 2014, Watson et al. 2017). Increasing the transparency of the supply chain is one of the biggest challenges in managing sustainable trade in marine species (Iles 2007, Pramod et al. 2014). In addition, I also found that key source countries for CITES marine species were often different from those dominating supply of other animal taxa. Those countries need to develop strong management approaches for marine species – and these may not be the same as for terrestrial taxa. For example, previous study has suggested that conserving the larger individuals of a species and protecting habitats at large geographic scale are particular important for marine species (McClenachan et al. 2016).   By identifying the important source and import countries across taxa, this study provides insights in prioritizing conservation efforts at a global scale for multiple species (Kremen et al. 2008). Previous studies on the management of CITES species mainly focus on single taxon (Bruckner 2001, Carpenter et al. 2005, 2014), overlooking the fact that national governments (including   120 management agencies and Customs) are responsible for multiple species. When the number of species in their responsibility increases, more financial and technical support are needed for a country to achieve its CITES requirements. My cross-species analyses are therefore useful for the CITES Secretariat to identify the countries that may need increased support.   6.3 Implications  My thesis provides a pioneering thorough examination on the trade in marine fishes under a global evolution of governance philosophy for fisheries management. The paradigm for fisheries management has shifted several times in history, and notable changes include a switch from maximizing production to sustainable fishing, from single species management to ecosystem based management, and from national management to regional/global planning (Castilla and Defeo 2005, Hughes et al. 2005, Pramod et al. 2014). One of the new paradigms for marine fisheries management is treating marine fishes as wildlife, with an explicit goal of managing fisheries is to conserve the species. Therefore, the applicability of international wildlife trade policy for the conservation of marine fishes needs to be considered (Caddy 1999, Vincent et al. 2013).   Through my thesis, I show that regulating exports of wildlife was associated with changes in seahorse trade through national management and behavior changes in traders. However, the challenges that have appeared in the management of non-marine wildlife trade, such as illegal trade, weak enforcement, and lack of/delay in deployment of extraction measures, were also common for marine fishes. The new paradigm is becoming more acceptable to policy makers as the international community increasingly shows commitment towards marine conservation as to   121 other wildlife. Examples of this commitment are found in managers’ consideration of IUCN Red List assessments for marine fishes in fisheries management and through the inclusion of marine species under the purviews of international agreements (Lascelles et al. 2014, Vincent and Foster 2017). However, more commercial marine species need to receive similar conservation attention, not least because FAO regimes for those species are commonly still guided more by economic values rather by sustainability norms (Lobo and Jacques 2017).  My research on seahorses has implications for the management of other marine fishes and Appendix II species. I found the trade in seahorses changed after CITES regulations in terms of volume, prices, and source countries, highlighting the importance of monitoring the trade (Raymakers and Hoover 2002, Poole and Shepherd 2016). Such changes could have been caused by CITES or by other factors (e.g. changes in demand or population size) that should be taken into account for adaptive management. I emphasize that the effects of policy instruments on trade and extraction is important in making NDFs, and should be incorporated as part of risk analysis and scenario testing for marine fishes (Smith et al. 2011). In exploring the dynamics in trade, my study demonstrated how to include information from various resources to complement the CITES records.  Although my research focuses on the effects of CITES, the results of my study are relevant to the implementation of other multilateral environmental treaties. Countries that I identified as encountering challenges in implementing CITES might also encounter difficulties in fulfilling other international environmental agreements. The goal of CITES (ensuring that exports do not threaten survival of wild populations) requires Parties to ensure sustainable extraction of the   122 species (CITES 1973), just as in RFMOs and other treaties (e.g., CBD) (Abensperg-Traun 2009, Vincent et al. 2013). CITES regulations serve as a complementary policy instrument to fisheries management, and countries should regard these different agreements as fostering a combination of policy tools to approach a complex management/conservation challenge (Young 2011).    6.4 Limitations   My study found that the quantitative data obtained from local knowledge, while providing valuable initial information for data-poor species, contains uncertainty because of recall bias and respondent attitudes. In addition to the recall biases identified in my Chapter 1, I note that other biases may exist in the fishers’ reports. For example, the fishers and traders may tend to under-report their catch and trade, considering that seahorses are species with conservation concerns (Filion 1981, Jones et al. 2008, Rist et al. 2010). These unavoidable biases in the interview data indicate that we should treat the information carefully and corroborate the results with other data sources, most of which have their own biases.  My research using official data (Customs and CITES records) does not capture the illegal, unreported and unregulated wildlife trade. As a result, the actual trade volume may be higher than I discerned, with more countries involved than the official data shows (Bruckner 2001, Lam et al. 2014, Foster et al. 2016). However, my findings in Chapter 5 on the changes in the relative species richness in trade and countries’ connections for each country should be legitimate, under the assumption that a country’s reporting rates and reliability did not change over time.      123 In the analyses on CITES data across taxa, I focused on species richness and countries’ connections, without addressing the trade volumes. This may lead to different results of determining the “key countries” in wildlife trade, since the number of species in trade often does not correspond to the trade volume (Goss and Cumming 2013). Nevertheless, the number of species in trade and countries’ connections are appropriate indicators for conservation planning for multiple species, because they are additive across species with different biological traits.  In contrast, trade volumes (e.g., number of individuals) are more comparable for species with similar reproduction rates and initial population size.   6.5 Future Directions Based on the findings of my thesis, I identified several directions for research and management that might produce a better conservation outcomes for seahorses, small marine bycatch species, species listed in CITES Appendix II, and all wildlife in trade.  6.5.1 Reducing market demand for conservation of seahorses caught incidentally Although there are concerns that trade regulations or reduction in demand may do little to reduce extraction of bycatch species (e.g., seahorses), I argue that weakening the market demand might increase the probability that bycaught seahorses would be released. Based on my personal observations in Thailand, most seahorses were landed alive in small-scale fisheries and trawl fisheries with short duration hauls or fishing trips. A survey in India also found that small-scale fishers in the states that have little seahorse trade tended to release the seahorses they caught, while fishers in the trade-intensive regions tended to keep and sell the fishes (T. Vaidyanathan, pers. comm., 2017). This implies that reduction in demand has potential to reduce seahorse   124 extraction, although survival rates of released seahorses are unknown. This is particularly important since small-scale fisheries (e.g., scoop nets and gillnets) contributed more than half of the seahorse bycatch globally (Lawson et al. 2017). More studies are needed to verify the conditions under which fishers release bycatch, and to explore possible modification of gear and methods to improve survival rates of released seahorses (and associated economic implications).  6.5.2 Management of bycatch in trawl fisheries Globally, seahorse conservation would be hugely advanced by reducing bycatch of non-selective gears. The proportion of bycatch in trawl fisheries reaches up to 90% of the total catch, with thousands of species involved (Alverson et al. 1994). In many places, most of the bycatch from trawl fisheries are reduced for animal/fish feed or fish oil, or shredded for surimi (Edwards et al. 2004, Nunoo et al. 2009). Any economic value of non-target species, even if slight, can help lower fishers’ motivation to reduce bycatch (and thus pressure on wild populations). Reducing the damaging effects of trawl fisheries generally depends on restricting fishing effort, constraining trawling spatially and temporally, and modifying gears (e.g., the use of bycatch reduction devices) (Broadhurst 2000). However, studies on reducing bycatch in trawl fisheries have hitherto mainly focused on large marine animals, such as sea turtles (Baum et al. 2003b, Lewison et al. 2004). While RFMOs have increased their attentions to the bycatch issues, those measures are either sporadic and lack of enforcement (Gilman et al. 2014). My work points to the even more substantial challenge of reducing trawl pressures on thousands of small species. This challenge will not be addressed by gear modifications, making trawl exclusion zones central and vital elements of any reconciliation of conservation with fisheries.       125 6.5.3 Prioritizing conservation efforts among CITES listed species  A fundamental question for conservation of species and spaces is how to allocate conservation efforts in the most cost-effective way. The CITES database, which comprises the most abundant trade information for more than 35,000 species, can be used to investigate such question for the species regulated by CITES. Specifically, we can explore two questions with the assistance of CITES data: (1) how to design management plans based on a higher taxonomic level if the species level data is unavailable or unreliable; and (2) where should we focus conservation investment at a global scale.   First, the information recorded in CITES database can help us to examine how likely we can design management plans based on a higher taxonomic level, if species-level data is unavailable or unreliable. Ideally, species-level data can provide the most detailed information for conservation and trade management. However, identifying the specimens in trade to species level can be difficult for some organisms (e.g., corals), and training officers in the management agency and Customs for such identification can be costly. If we are able to find a higher taxon surrogacy to use for predicting the patterns at species level, conservation plans based on a higher taxonomy level will be less affected by the lack of data or misidentification at species level. To do so, we can investigate if the trade patterns of a taxon differed by the taxonomy level used in analysis (e.g., Family, Genus, or Species). While species richness has been found strongly correlated with genera and family richness spatially and temporally for many taxa (e.g., stream macro-invertebrates, diatoms, and birds) (Heino and Soininen 2007, Kallimanis et al. 2012), it will be interesting and useful to know if such correlation exists in trade data and whether we can plan the management strategies at a higher taxonomic level.    126  Second, combining CITES trade data with other ecological information about a species or taxon can help prioritize conservation efforts in countries that play the most active roles in trade. Previous studies on conservation investment at global scale focused on regions/countries with the highest species richness/biodiversity. However, it has been controversial as to whether the “species richness” should refer to general species richness, endemic species richness, endangered species richness, or other indicators (Ceballos et al. 2005, Grenyer et al. 2006, Bode et al. 2008). In Chapter 5 of my thesis, I examined the species richness in exports/imports to identify the countries with a high number of native species in trade. However, we should also consider the number of rare species (e.g., endemic species) in trade, the ecosystem service each species can provide (Naidoo et al. 2008), the conservation status of each species (e.g., determining by IUCN risk assessment), and the costs of implementing conservation actions in each country (Bode et al. 2008).   6.5.4 Addressing illegal wildlife trade Reaching beyond the scope of my study, there is an urgent need to address illegal wildlife trade, though such information is patchy and sparse. In such situation, how do we best use the legal trade data and incorporate the available illegal wildlife trade information to understand the trade? To do so, we need to investigate the relationship between legal and illegal wildlife trade. A study conducted by Patel (2015) found that legal wildlife trade data can predict 40-80% of the countries involved in illegal wildlife trade for a specific product type for rhinos, tigers, and elephants. However, it is unknown if such predictions are consistent for other wildlife in trade. On the other hand, there are only a few assessments for how much of the total trade   127 volume/value is reflected in the legal trade records (Song 2008, Patel 2015). Considering the obscure nature of wildlife trade, incorporating all available data, including legal records and reports on illegal trade, is critical in decision-making.   6.5.5 Enhancing the role of industry in fostering sustainable wildlife trade In addition to international and national regimes, we need action by industry to help achieve sustainable trade. We should encourage the traders to provide their perspectives in the best way to ensure their commodities are sustainable sourced. For example, a minimum size limit for all seahorse species has been proposed for enhancing the sustainable seahorse trade (Foster and Vincent 2005). Seahorse traders have agreed that such size limit is appropriate for both biological and economic sustainability, and is easy to comply (Martin-Smith et al. 2004). Another example is the voluntary bans on the transport of shark fins and endangered wildlife trophies adopted by flight and shipping companies. However, if and such bans are effective has not yet been investigated. Strengthening cross-national, bottom-up approaches from industry may complement the discordance in the top-down management among national governments.    6.6 Final remarks My research reveals the immense potential of international trade regulations for the management of marine fishes. While my research focuses on seahorses, I am intrigued about the prospects of trade measures for other commercially important marine species. As for wildlife trade, illegal trade is always a huge problem for effective management. It is always a challenge to enhance the traceability of the supply chain, but I expect that this will become increasingly possible in this era where many trade transactions are happening online. I foresee that the next step in wildlife   128 trade research is to understand online trading behaviours, either on social media or on shopping websites, and design management for a trade that is no longer restricted by geographical distances and boundaries.   129 Bibliography Aarts, G., and Poos, J.J. 2009. Comprehensive discard reconstruction and abundance estimation using flexible selectivity functions. ICES J. Mar. Sci. J. du Cons. 66(4): 763–771. doi:10.1093/icesjms/fsp033. Abensperg-Traun, M. 2009. CITES, sustainable use of wild species and incentive-driven conservation in developing countries, with an emphasis on southern Africa. Biol. Conserv. 142: 948–963. Available from http://www.rhinoresourcecenter.com/pdf_files/125/1255713712.pdf [accessed 13 January 2016]. Agnew, D.J., Pearce, J., Pramod, G., Peatman, T., Watson, R., Beddington, J.R., and Pitcher, T.J. 2009. Estimating the worldwide extent of illegal fishing. PLoS One 4(2): e4570. doi:10.1371/journal.pone.0004570. Ahmed, M., Boonchuwongse, P., Dechboon, W., and Squires, D. 2007. Overfishing in the Gulf of Thailand: policy challenges and bioeconomic analysis. Environ. Dev. Econ. 12(1): 145. Cambridge University Press. doi:10.1017/S1355770X06003433. Aiken, K.A., Andre Kong, G., Smikle, S., Mahon, R., and Appeldoorn, R. 1999. The queen conch fishery on Pedro Bank, Jamaica: discovery, development, management. Ocean Coast. Manag. 42(12): 1069–1081. doi:10.1016/S0964-5691(99)00063-0. Alverson, D.L., Murawski, S.A., and Pope, L.G. 1994. A global assessment of fisheries bycatch and discards. Rome. Available from http://www.fao.org/docrep/003/t4890e/t4890e00.htm [accessed 25 January 2016]. Anders, S.M., and Caswell, J.A. 2009. Standards-as-Barriers versus Standards-as-Catalysts: Assessing the Impact of HACCP Implementation on U.S. Seafood Imports. Am. J. Agric.   130 Econ. 91(2): 310–321. Available from http://ajae.oxfordjournals.org/content/91/2/310.short [accessed 15 December 2016]. Anderson, S., Flemming, J., Watson, R., and Lotze, H. 2011. Serial exploitation of global sea cucumber fisheries. Fish Fish. 12(3): 317–339. Available from http://seananderson.ca/papers/Anderson_etal_2011_seacucumbers_with_supplement.pdf [accessed 9 June 2015]. Andrew, N., Agatsuma, Y., and Ballesteros, E. 2003. Status and management of world sea urchin fisheries. Oceanogr. Mar. Biol. Annu. Rev. 40: 343–425. Available from https://books.google.ca/books?hl=en&lr=&id=2H4aydD3VWgC&oi=fnd&pg=PA343&dq=Status+and+management+of+world+sea+urchin+fisheries&ots=C-A5OrLpz6&sig=ysvO8SQkYmN5Nu-mdE1Sx20iXJU [accessed 1 February 2017]. Ardron, J.A., Rayfuse, R., Gjerde, K., and Warner, R. 2014. The sustainable use and conservation of biodiversity in ABNJ: What can be achieved using existing international agreements? Mar. Policy 49: 98–108. doi:10.1016/j.marpol.2014.02.011. Arlinghaus, R., and Cooke, S. 2009. Recreational Fisheries: Socioeconomic Importance, Conservation Issues and Management Challenges. In Recreational Hunting, Conservation and Rural Livelihoods: Science and Practice. Edited by B. Dickson, J. Hutton, and W.M. Adams. Blackwell Publishing Ltd. pp. 39–58. Available from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.456.6238&rep=rep1&type=pdf#page=53 [accessed 15 May 2017]. Auliya, M., Altherr, S., Ariano-Sanchez, D., Baard, E.H., Brown, C., Brown, R.M., Cantu, J.C., Gentile, G., Gildenhuys, P., Henningheim, E., Hintzmann, J., Kanari, K., Krvavac, M., Lettink, M., Lippert, J., Luiselli, L., Nilson, G., Nguyen, T.Q., Nijman, V., Parham, J.F.,   131 Pasachnik, S.A., Pedrono, M., Rauhaus, A., Córdova, D.R., Sanchez, M.E., Schepp, U., van Schingen, M., Schneeweiss, N., Segniagbeto, G.H., Somaweera, R., Sy, E.Y., Türkozan, O., Vinke, S., Vinke, T., Vyas, R., Williamson, S., and Ziegler, T. 2016. Trade in live reptiles, its impact on wild populations, and the role of the European market. Biol. Conserv. 204: 103–119. doi:10.1016/j.biocon.2016.05.017. Babcock, E.A., and Pikitch, E.K. 2011. A dynamic programming model of fishing strategy choice in a multispecies trawl fishery with trip limits. Can. J. Fish. Aquat. Sci. 57(2): 357–370. NRC Research Press Ottawa, Canada. Available from http://www.nrcresearchpress.com/doi/abs/10.1139/f99-257#.Vt-Nu5MrLkJ [accessed 9 March 2016]. Baier, S.L., and Bergstrand, J.H. 2007. Do free trade agreements actually increase members’ international trade? J. Int. Econ. 71(1): 72–95. doi:10.1016/j.jinteco.2006.02.005. Ban, N.C., Hansen, G.J.A., Jones, M., and Vincent, A.C.J. 2009. Systematic marine conservation planning in data-poor regions: Socioeconomic data is essential. Mar. Policy 33(5): 794–800. doi:10.1016/j.marpol.2009.02.011. Barbeaux, S.J., Horne, J.K., and Ianelli, J.N. 2014. A novel approach for estimating location and scale specific fishing exploitation rates of eastern Bering Sea walleye pollock (Theragra chalcogramma). Fish. Res. 153(October 2015): 69–82. doi:10.1016/j.fishres.2014.01.004. Barbier, E.B., Bockstael, N., Burgess, J.C., and Strand, I. 1995. The linkages between the timber trade and tropical deforestation?- Indonesia. World Econ. 18(3): 411–442. Blackwell Publishing Ltd. doi:10.1111/j.1467-9701.1995.tb00222.x. Barden, A., Anak, N., Mulliken, T., and Song, M. 2000. Heart of the matter: agarwood use and trade and CITES implementation for Aquilaria malaccensis. In TRAFFIC. Available from   132 https://www.traffic.org/forestry-reports/traffic_pub_forestry7.pdf [accessed 31 March 2017]. Barden, C. 1994. Combating the illegal timber trade—is there a role for ITTO? Green globe Yearb. Int. co-operation. Available from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.730.1317&rep=rep1&type=pdf [accessed 26 January 2017]. Baum, J., Meeuwig, J., and Vincent, A. 2003a. Bycatch of lined seahorses (Hippocampus erectus) in a Gulf of Mexico shrimp trawl fishery. Fish. Bull. 101(4): 721–731. Available from http://aquaticcommons.org/15162/ [accessed 15 November 2015]. Baum, J., Myers, R., Kehler, D., and Worm, B. 2003b. Collapse and conservation of shark populations in the Northwest Atlantic. Available from http://science.sciencemag.org/content/299/5605/389.short [accessed 15 May 2017]. Baum, J.K., and Vincent, A.C.J. 2005a. Magnitude and inferred impacts of the seahorse trade in Latin America. Environ. Conserv. 32(4): 305–319. doi:10.1017/S0376892905002481. Baum, J.K., and Vincent, A.C.J. 2005b. Magnitude and inferred impacts of the seahorse trade in Latin America. Environ. Conserv. 32(4): 305. Cambridge University Press. doi:10.1017/S0376892905002481. Baylis, K., Nogueira, L., and Pace, K. 2011. Food import refusals: rvidence from the European Union. Am. J. Agric. Econ. 93(2): 566–572. doi:10.1093/ajae/aaql49. Beaman, J., Vaske, J.J., and Miller, C.A. 2005. Cognitive processes in hunters’ recall of participation and harvest estimates. J. Wildl. Manage. 69(3): 967–975. The Wildlife Society. doi:10.2193/0022-541X(2005)069[0967:CPIHRO]2.0.CO;2.   133 Beaudreau, A.H., and Levin, P.S. 2014. Advancing the use of local ecological knowledge for assessing data-poor species in coastal ecosystems. Ecol. Appl. 24(2): 244–56. doi:10.1890/13-0817.1. Beissinger, S.R., and Bucher., E.H. 1992. Can parrots be conserved through sustainable harvesting? Bioscience 42(3): 164–173. Available from http://www.jstor.org/stable/1311821?seq=1#page_scan_tab_contents [accessed 11 April 2016]. Bennett, E.L. 2015. Legal ivory trade in a corrupt world and its impact on African elephant populations. Conserv. Biol. 29(1): 54–60. doi:10.1111/cobi.12377. Bennett, N., and Dearden, P. 2014. Why local people do not support conservation: community perceptions of marine protected area livelihood impacts, governance and management in Thailand. Mar. Policy. Available from http://www.sciencedirect.com/science/article/pii/S0308597X13001711 [accessed 15 November 2015]. Bennett, N., Dearden, P., Murray, G., and Kadfak, A. 2014. The capacity to adapt?: communities in a changing climate, environment, and economy on the northern Andaman coast of Thailand. Ecol. Soc. 19(2): 5. Available from https://www.researchgate.net/profile/Nathan_Bennett3/publication/261546911_The_capacity_to_adapt_Communities_in_a_changing_climate_environment_and_economy_on_the_northern_Andaman_coast_of_Thailand/links/0046353486c06e239e000000.pdf [accessed 11 April 2016]. Berkes, F. 2006. Globalization, Roving Bandits, and Marine Resources. Science (80-. ). 311(5767): 1557–1558. doi:10.1126/science.1122804.   134 Berkes, F., Colding, J., and Folke, C. 2000. Rediscovery of traditional ecological knowledge as adaptive management. Ecol. Appl. 10(5): 1251–1262. doi:10.1890/1051-0761(2000)010[1251:ROTEKA]2.0.CO;2. Biernacki, P., and Waldorf, D. 1981. Snowball sampling: Problems and techniques of chain referral sampling. Sociol. Methods Res. 10(2): 141–163. Available from http://smr.sagepub.com/content/10/2/141.short [accessed 26 May 2016]. Blundell, A.G., and Mascia, M.B. 2005. Discrepancies in reported levels of international wildlife trade. Conserv. Biol. 19: 2020–2025. doi:10.1111/j.1523-1739.2005.00253.x. Bode, M., Wilson, K.A., Brooks, T.M., Turner, W.R., Mittermeier, R.A., McBride, M.F., Underwood, E.C., and Possingham, H.P. 2008. Cost-effective global conservation spending is robust to taxonomic group. Proc. Natl. Acad. Sci. U. S. A. 105(17): 6498–501. National Academy of Sciences. doi:10.1073/pnas.0710705105. Bosshard, P. 2008. China’s Environmental Footprint in Africa. China Africa Policy Brief. (3). Available from http://198.61.250.55/sites/default/files/China’s-Environmental-Footprint-in-Africa.pdf [accessed 26 January 2017]. Branch, T.A., Jensen, O.P., Ricard, D., Ye, Y., and Hilborn, R. 2011. Contrasting global trends in marine fishery status obtained from catches and from stock assessments. Conserv. Biol. 25(4): 777–86. doi:10.1111/j.1523-1739.2011.01687.x. Branch, T.A., Lobo, A.S., and Purcell, S.W. 2013. Opportunistic exploitation: an overlooked pathway to extinction. Trends Ecol. Evol. 28(7): 409–413. doi:10.1016/j.tree.2013.03.003. Brashares, J., Arcese, P., Sam, M., and Coppolillo, P. 2004. Bushmeat hunting, wildlife declines, and fish supply in West Africa. Available from http://science.sciencemag.org/content/306/5699/1180.short [accessed 26 April 2017].   135 Broadhurst, M.K. 2000. Modifications to reduce bycatch in prawn trawls: A review and framework for development. Rev. Fish Biol. Fish. 10(1): 27–60. Kluwer Academic Publishers. doi:10.1023/A:1008936820089. Brooks, T.M., Mittermeier, R.A., Da Fonseca, G.A.B., Gerlach, J., Hoffmann, M., Lamoreux, J.F., Mittermeier, C.G., Pilgrim, J.D., and Rodrigues, A.S.L. 2006. Global Biodiversity Conservation Priorities. 58. doi:10.1126/science.1127609. Brooks, T.M., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Oldfield, S., Magin, G., and Hilton-Taylor, C. 2002. Habitat Loss and Extinction in the Hotspots of Biodiversity. Conserv. Biol. 16(4): 909–923. Blackwell Science Inc. doi:10.1046/j.1523-1739.2002.00530.x. Bruckner, A.W. 2001. Tracking the trade in ornamental coral reef organisms: The importance of CITES and its limitations. Aquarium Sci. Conserv. 3: 79–94. doi:10.1023/A:1011369015080. Burton, S., and Blair, E. 1991. Task conditions, response formulation processes, and response accuracy for behavioral frequency questions in surveys. Public Opin. Q. Available from https://academic.oup.com/poq/article-abstract/55/1/50/1819914 [accessed 18 August 2017]. Bush, E.R., Baker, S.E., and Macdonald, D.W. 2014. Global trade in exotic pets 2006-2012. Conserv. Biol. 28(3): 663–676. doi:10.1111/cobi.12240. Byg, A., Theilade, I., Nielsen, M., and Lund, J. 2012. Local ecological knowledge and its relevance for management and research. Dev. Briefs. Policy 16. Available from http://curis.ku.dk/ws/files/37364425/Local_ecological_knowledge.pdf [accessed 14 November 2015].   136 Caddy, J.F. 1999. Fisheries management in the twenty-first century: will new paradigms apply? Rev. Fish Biol. Fish. 9(1): 1–43. doi:10.1023/A:1008829909601. Cahill, A., Walker, J., and Marsden, S. 2006. Recovery within a population of the Critically Endangered citron-crested cockatoo Cacatua sulphurea citrinocristata in Indonesia after 10 years of international trade. Oryx. Available from http://search.proquest.com/openview/8438b7c8a26b8be632fc0ac0f4e642af/1?pq-origsite=gscholar&cbl=37514 [accessed 25 April 2017]. Cannell, C., Oksenberg, L., and Converse, J. 1977. Striving for response accuracy: experiments in new interviewing techniques. J. Mark. Res.: 306–315. Available from http://www.jstor.org/stable/3150768 [accessed 25 November 2015]. Carpenter, A.I. 2004. The dynamics of the global trade in chameleons. Biol. Conserv. 120(2): 291–301. doi:10.1016/j.biocon.2004.03.002. Carpenter, A.I., Andreone, F., Moore, R.D., and Griffiths, R. a. 2014. A review of the international trade in amphibians: the types, levels and dynamics of trade in CITES-listed species. Oryx 48(February 2012): 1–10. doi:10.1017/S0030605312001627. Carpenter, A.I., Robson, O., Rowcliffe, J.M., and Watkinson, A.R. 2005. The impacts of international and national governance changes on a traded resource: A case study of Madagascar and its chameleon trade. Biol. Conserv. 123(3): 279–287. doi:10.1016/j.biocon.2004.11.015. Carr, L.M., and Heyman, W.D. 2012. “It’s about seeing what’s actually out there”: Quantifying fishers’ ecological knowledge and biases in a small-scale commercial fishery as a path toward co-management. Ocean Coast. Manag. 69: 118–132. Elsevier Ltd. doi:10.1016/j.ocecoaman.2012.07.018.   137 Casale, P. 2010. Sea turtles in the Mediterranean: distribution, threats and conservation priorities. IUCN. Available from https://books.google.ca/books?hl=en&lr=&id=hY3HgFUa0JQC&oi=fnd&pg=PP1&dq=Sea+turtles+in+the+Mediterranean+:+Distribution,+threats+and+conservation+priorities&ots=VFQSbcrTIi&sig=NwSkgL7XIHaKFFTTUJYfRms-Pes [accessed 28 April 2016]. Castilla, J.C., and Defeo, O. 2005. Paradigm shifts needed for world fisheries. Science (80-. ). 309(5739): 1324. doi:10.1126/science.309.5739.1324c. Catarci, C. 2004. World markets and industry of selected commercially-exploited aquatic species with an international conservation profile. Food and Agriculture Organization of the United Nations. Available from http://www.sidalc.net/cgi-bin/wxis.exe/?IsisScript=LIBRI.xis&method=post&formato=2&cantidad=1&expresion=mfn=019332 [accessed 28 April 2016]. Ceballos, G., Ehrlich, P., Soberón, J., and Salazar, I. 2005. Global mammal conservation: what must we manage? Science (80-. ). Available from http://science.sciencemag.org/content/309/5734/603.short [accessed 9 May 2017]. Challender, D.W.S., Harrop, S.R., and MacMillan, D.C. 2015a. Towards informed and multi-faceted wildlife trade interventions. Glob. Ecol. Conserv. 3: 129–148. doi:10.1016/j.gecco.2014.11.010. Challender, D.W.S., Harrop, S.R., and MacMillan, D.C. 2015b. Understanding markets to conserve trade-threatened species in CITES. Biol. Conserv. 187: 249–259. Elsevier Ltd. doi:10.1016/j.biocon.2015.04.015.   138 Chen, T.-H., Chang, H.-C., and Lue, K.-Y. 2009. Unregulated Trade in Turtle Shells for Chinese Traditional Medicine in East and Southeast Asia: The Case of Taiwan. Chelonian Conserv. Biol. 8(1): 11–18.  Chelonian Research Foundation . doi:10.2744/CCB-0747.1. Chichilnisky, G. 1993. North-South trade and the dynamics of renewable resources. Struct. Chang. Econ. Dyn. 4(2): 219–248. doi:10.1016/0954-349X(93)90017-E. Choo, C., and Liew, H. 2005. Exploitation and trade in seahorses in Peninsular Malaysia. Malayan Nat. J. 57(1): 57–66. Available from http://www.researchgate.net/profile/Hock_Chark_Liew/publication/237036146_Exploitation_and_Trade_in_Seahorses_in_Peninsular_Malaysia/links/0deec51af5465264df000000.pdf [accessed 15 November 2015]. CITES. 1973. Convention on International Trade in Endangered Species of Wild Fauna and Flora. doi:10.1016/S0378-777X(84)80087-6. CITES. 2004. Conf. 9.24 (Rev. CoP13) Criteria for amendment of Appendix I and II. Available from https://cites.org/eng/res/all/09/E09-24R13.pdf [accessed 25 April 2017]. CITES. 2012. Review of Significant Trade: Species selected by the CITES Animals Committee following CoP14 and retained in the review following AC25. Available from https://cites.org/eng/com/ac/26/E26-12-02-A.pdf [accessed 25 January 2016]. CITES. 2013. Building in-country capacity to undertake Non-Detriment Findings with regard to Hippocampus species in Indonesia, Thailand and Viet Nam. Geneva, Switzerland. CITES. 2016a. SC67 Doc.15 Additional information on seahorses in Thailand and Temporary ceasing the issuance of export permits for seahorses (Hippocampus spp.). Available from https://cites.org/sites/default/files/eng/com/sc/67/E-SC67-15-A2.pdf [accessed 25 April 2017].   139 CITES. 2016b. Notification to the Parties No. 2016/018. Available from https://cites.org/sites/default/files/notif/E-Notif-2016-018.pdf [accessed 14 May 2016]. CITES. 2017. Countries currently subject to a recommendation to suspend trade. Available from https://cites.org/eng/resources/ref/suspend.php [accessed 20 March 2017]. CITES AC26. 2012. Review of significant trade in specimens of Appendix II species. CITES Resolution Conference 16.7. 2013. Non-detrimental Findings. Available from https://cites.org/sites/default/files/eng/res/16/E-Res-16-07.pdf [accessed 15 June 2016]. CITES Secretariat. 2013. Notification to the Parties No. 2013/017. In CITES. Available from https://cites.org/sites/default/files/eng/notif/2013/E-Notif-2013-017.pdf [accessed 20 March 2017]. CITES Secretariat. 2016. SC66 Doc. 31.1 Review of Significant Trade in specimens of Appendix-II species. Geneva, Switzerland. Available from https://cites.org/sites/default/files/eng/com/sc/66/E-SC66-31-01.pdf. Cochrane, K. 2015. Use and misuse of CITES as a management tool for commercially-exploited aquatic species. Mar. Policy 59: 16–31. Elsevier. doi:10.1016/j.marpol.2015.04.015. Connelly, N.A., and Brown, T.L. 1995. Use of angler diaries to examine biases associated with 12-Month recall on mail questionnaires. Trans. Am. Fish. Soc. 124(3): 413–422. Taylor & Francis Group. doi:10.1577/1548-8659(1995)124<0413:UOADTE>2.3.CO;2. Cooney, R., and Abensperg-Traun, M. 2013. Raising Local Community Voices: CITES, Livelihoods and Sustainable Use. Rev. Eur. Comp. Int. Environ. Law 22(3): 301–310. doi:10.1111/reel.12038. Cooney, R., and Jepson, P. 2006. The international wild bird trade: what’s wrong with blanket bans? Oryx 40(1): 18–23. Available from   140 http://www.geog.ox.ac.uk/research/publications/rcooney-pjepson-oryx-2006.pdf [accessed 15 January 2016]. Costello, C., Ovando, D., Hilborn, R., Gaines, S.D., Deschenes, O., and Lester, S.E. 2012. Status and solutions for the world’s unassessed fisheries. Science 338(6106): 517–20. doi:10.1126/science.1223389. Coughlin, S.S. 1990. Recall bias in epidemiologic studies. J. Clin. Epidemiol. 43(1): 87–91. doi:10.1016/0895-4356(90)90060-3. Courchamp, F., Angulo, E., Rivalan, P., Hall, R.J., Signoret, L., Bull, L., and Meinard, Y. 2006. Rarity value and species extinction: the anthropogenic Allee effect. PLoS Biol. 4(12): e415. Public Library of Science. doi:10.1371/journal.pbio.0040415. Crona, B.I., Daw, T.M., Swartz, W., Norström, A. V., Nyström, M., Thyresson, M., Folke, C., Hentati-Sundberg, J., Österblom, H., Deutsch, L., and Troell, M. 2015. Masked, diluted and drowned out: How global seafood trade weakens signals from marine ecosystems. Fish Fish. 17(17): 1175–1182. doi:10.1111/faf.12109. Csárdi, G., and Nepusz, T. 2006. The igraph software package for complex network research. InterJournal Complex Syst. 1695: 1695. Available from https://pdfs.semanticscholar.org/1d27/44b83519657f5f2610698a8ddd177ced4f5c.pdf [accessed 7 April 2017]. D’Cruze, N., and Macdonald, D.W. 2015. Clouded in mystery: the global trade in clouded leopards. Biodivers. Conserv. 24(14): 3505–3526. Springer Netherlands. doi:10.1007/s10531-015-1010-9.   141 Dascal, D., Mattas, K., and Tzouvelekas, V. 2002. An analysis of EU wine trade: A gravity model approach. Int. Adv. Econ. Res. 8(2): 135–147. Kluwer Academic Publishers. doi:10.1007/BF02295344. Davidson, L.N.K., Krawchuk, M.A., and Dulvy, N.K. 2016. Why have global shark and ray landings declined: Improved management or overfishing? Fish Fish. 17(2): 438–458. doi:10.1111/faf.12119. Daw, T.M. 2008. How Fishers Count. Enganging with fishers knooldege in fisheries science and management. Newcastle. Available from https://theses.ncl.ac.uk/dspace/handle/10443/114 [accessed 3 May 2017]. Daw, T.M., Robinson, J., and Graham, N. a. J. 2011. Perceptions of trends in Seychelles artisanal trap fisheries: comparing catch monitoring, underwater visual census and fishers’ knowledge. Environ. Conserv. 38(1): 75–88. doi:10.1017/S0376892910000901. Denzin, N., and Lincoln, Y. 2011. The SAGE handbook of qualitative research. Available from https://books.google.ca/books?hl=en&lr=&id=AIRpMHgBYqIC&oi=fnd&pg=PR1&dq=denzin+lincoln+&ots=knDRBGfCgd&sig=MCYjiTKZ5SemY5lp6pc_9Tk7vto [accessed 15 November 2015]. DG MARE. 2014. Study on the state of play regarding application and implementation of council regulation (EC) NO 1005/2008 of 29 September 2008, establishing a community system to prevent, deter and eliminate illegal, unreported and unregulated fishing (IUU Regulation). Dickson, B. 2013. Endangered Species Threatened Convention: The Past, Present and Future of CITES, the Convention on International Trade in Endangered Species of Wild Fauna and Flora. Taylor & Francis. Available from   142 https://books.google.com/books?hl=en&lr=&id=iUD5AQAAQBAJ&pgis=1 [accessed 24 April 2016]. DoF. 2015. Fisher Census. Available from http://www.platalay.com/boatsurvey2554/index.php [accessed 30 December 2015]. Dulvy, N., Baum, J., Clarke, S., and Domingo, A. 2008. You can swim but you can’t hide: the global status and conservation of oceanic pelagic sharks and rays. Aquat. Conserv. Mar. Freshw. Ecosyst. 18: 459–482. Available from http://www.researchgate.net/profile/Nicholas_Dulvy/publication/237397030_You_can_swim_but_you_can’t_hide_the_global_status_and_conservation_of_oceanic_pelagic_sharks_and_rays/links/547cbe3f0cf285ad5b088568.pdf [accessed 26 January 2016]. Edwards, P., Edwards, P., Allan, G.L., and Allan, G.L. 2004. A survey of marine trash fish and fish meal as aquaculture feed ingredients in Vietnam. In Australian Centre for International Agricultural Research. Available from http://leanhtuan.vnweblogs.com/gallery/23458/03 - trash fish report - vietnam aciar.pdf [accessed 10 May 2017]. Efron, B. 1979. Bootstrap methods: Another look at the Jackknife. Ann. Stat. 7(1): 1–26. Institute of Mathematical Statistics. doi:10.1214/aos/1176344552. Engler, M. 2008. The value of international wildlife trade. TRAFFIC Bull. 22(1): 4–5. Eriksson, H., and Clarke, S. 2015. Chinese market responses to overexploitation of sharks and sea cucumbers. Biol. Conserv. 184: 163–173. doi:10.1016/j.biocon.2015.01.018. Erisman, B., Mascareñas-Osorio, I., López-Sagástegui, C., Moreno-Báez, M., Jiménez-Esquivel, V., and Aburto-Oropeza, O. 2015. A comparison of fishing activities between two coastal communities within a biosphere reserve in the Upper Gulf of California. Fish. Res. 164: 254–265. doi:10.1016/j.fishres.2014.12.011.   143 FAO. 2014. The State of World Fisheries and Aquaculture. Rome. FAO. 2016. The State of World Fisheries and Aquaculture 2016. doi:92-5-105177-1. Filion, F. 1981. Importance of question wording and response burden in hunter surveys. J. Wildl. Manage.: 873–882. Available from http://www.cabdirect.org/abstracts/19821895334.html [accessed 18 October 2015]. Fisher, W., Grambsch, A., Eisenhower, D., and Morganstein, D. 1991. Length of recall period and accuracy of estimates from the national survey of fishing, hunting, and wildlife-associated recreation. Am. Fish. Soc. Symp. 12: 367–374. Available from http://sedarweb.org/docs/wsupp/SEDAR24-RD63_Fisher1991.pdf [accessed 8 March 2016]. Foster, S., and Vincent, A. 2004. Life history and ecology of seahorses: implications for conservation and management. J. Fish Biol. 65(1): 1–61. Available from http://onlinelibrary.wiley.com/doi/10.1111/j.0022-1112.2004.00429.x/pdf [accessed 15 November 2015]. Foster, S., and Vincent, A. 2005. Enhancing sustainability of the international trade in seahorses with a single minimum size limit. Conserv. Biol. 19(4): 1044–1050. Available from http://onlinelibrary.wiley.com/doi/10.1111/j.1523-1739.2004.00157.x-i1/full [accessed 25 January 2016]. Foster, S.J., Wiswedel, S., and Vincent, A.C.J. 2016. Opportunities and challenges for analysis of wildlife trade using CITES data - seahorses as a case study. Aquat. Conserv. Mar. Freshw. Ecosyst. 26(1): 154–172. doi:10.1002/aqc.2493. Freeman, L. 1978. Centrality in social networks conceptual clarification. Soc. Networks 1(3): 215–239. Available from   144 http://www.sciencedirect.com/science/article/pii/0378873378900217 [accessed 8 April 2017]. Gavin, M.C., Solomon, J.N., and Blank, S.G. 2010. Measuring and Monitoring Illegal Use of Natural Resources\rMedición y Monitoreo del Uso Ilegal de Recursos Naturales. Conserv. Biol. 24(1): 89–100. doi:10.1111/j.1523-1739.2009.01387.x. Gibbons, J.W., Scott, D.E., Rya, A.J., B U H L M A N N, K.A., Acey, T.R., Tuberv, D., Metts, B.S., Greene, J.L., Mills, T., Leiden, Y., Poppy, S., and Winne, C.T. 2000. The Global Decline of Reptiles, Déjà Vu Amphibians. @BULLET Biosci. 50(653). Available from https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/bioscience/50/8/10.1641_0006-3568(2000)050%5B0653_TGDORD%5D2.0.CO;2/3/50-8-653.pdf?Expires=1492045320&Signature=A8IuBCaar6lDEvM6JsPTXPU5aqYP~w5JUQj1l60OVDPYQ357p3ahkqSIzE7U9ksSE2vgUVcT-NW3aa-yzmpEv10sed9ibNc52HoMXjtWrx5CKRPt0n~IwAa-MzlraU52yMvdfpNO1tEZOi5JuEPFPMlU4lB6PB3YhIeqz92fXyieaDBlWQqPjn~yjPNCYHxJ4qQ9tUdy7-OZt8UiPGexGGcxSQu-CDOiCYi25yUQbC17ibZ4gZm0vo22Gv1adC9O2iEzqfcsxa9XWwx8gS3oBCJpiumOJuq3rwiVY2mNeZqK-HgJUmVTBYcikmvJE5dV7dD5aS2Qp3BZ0j5dh0osig__&Key-Pair-Id=APKAIUCZBIA4LVPAVW3Q [accessed 8 April 2017]. Giles, B.G., Ky, T.S., Hoang, D.H., and Vincent, A.C.J. 2006. The catch and trade of seahorses in Vietnam. Biodivers. Conserv. 15(8): 2497–2513. doi:10.1007/s10531-005-2432-6.   145 Giljum, S., and Eisenmenger, N. 2004a. North-South trade and the distribution of environmental goods and burdens: a biophysical perspective. J. Environ.  …. Available from http://jed.sagepub.com/content/13/1/73.short [accessed 11 June 2015]. Giljum, S., and Eisenmenger, N. 2004b. North-South Trade and the Distribution of Environmental Goods and Burdens: a Biophysical Perspective. J. Environ. Dev. 13. doi:10.1177/1070496503260974. Gilman, E., Passfield, K., and Nakamura, K. 2014. Performance of regional fisheries management organizations: ecosystem-based governance of bycatch and discards. Fish Fish. 15(2): 327–351. doi:10.1111/faf.12021. Gini, C. 1921. Measurement of Inequality of Incomes. Econ. J. 31(121): 124. doi:10.2307/2223319. Golden, C.D., Wrangham, R.W., and Brashares, J.S. 2013. Assessing the accuracy of interviewed recall for rare, highly seasonal events: the case of wildlife consumption in Madagascar. Anim. Conserv. 16(6): 597–603. doi:10.1111/acv.12047. Gordon, H.S. 1954. The economic theory of a common-property resource: the fishery. Source J. Polit. Econ. 62(2): 124–142. Available from http://www.jstor.org/stable/1825571 [accessed 10 May 2017]. Goss, J.R., and Cumming, G.S. 2013. Networks of wildlife translocations in developing countries: an emerging conservation issue? Front. Ecol. Environ. 11(5): 243–250. Ecological Society of America. doi:10.1890/120213. Grant, S., and Berkes, F. 2007. Fisher knowledge as expert system: A case from the longline fishery of Grenada, the Eastern Caribbean. Fish. Res. 84(2): 162–170. doi:10.1016/j.fishres.2006.10.012.   146 Grenyer, R., Orme, C.D.L., Jackson, S.F., Thomas, G.H., Davies, R.G., Davies, T.J., Jones, K.E., Olson, V.A., Ridgely, R.S., Rasmussen, P.C., Ding, T.-S., Bennett, P.M., Blackburn, T.M., Gaston, K.J., Gittleman, J.L., and Owens, I.P.F. 2006. Global distribution and conservation of rare and threatened vertebrates. Nature 444(7115): 93–96. doi:10.1038/nature05237. Hagan, N., Neis, B., and Baird, I.G. (Editors). 2007. Fishers’ Knowledge in Fisheries Science and Management. In UNESCO. Available from http://www.academia.edu/download/35590757/Fishers_Knowledge_2014.pdf [accessed 3 May 2017]. Hall, R.J., Milner-Gulland, E.J., and Courchamp, F. 2008. Endangering the endangered: The effects of perceived rarity on species exploitation. Conserv. Lett. 1(2): 75–81. Blackwell Publishing Inc. doi:10.1111/j.1755-263X.2008.00013.x. Hamilton, R., and Walter, R. 1999. Indigenous ecological knowledge and its role in fisheries research design: A case study from Roviana Lagoon, Western Province, Solomon Islands. SPC Tradit. Mar. Resour. Manag. Knowl. Inf. Bull. 11: 13–25. Available from http://www.spc.int/digitallibrary/doc/fame/infobull/trad/11/trad11_13_hamilton.pdf [accessed 15 November 2015]. Han, S. 2013. Taxonomy, resources, utilization and conservation of sea horses in China. Guangxi Normal University. doi:10.1017/CBO9781107415324.004. Harrison, R.D., Sreekar, R., Brodie, J.F., Brook, S., Luskin, M., O’Kelly, H., Rao, M., Scheffers, B., and Velho, N. 2016. Impacts of hunting on tropical forests in Southeast Asia. Conserv. Biol. 30(5): 972–981. doi:10.1111/cobi.12785.   147 Head, K., and Mayer, T. 2010. Illusory border effects: distance mismeasurement inflates estimates of home bias in trade. In The gravity model in international trade: Advances and applications. CEPII research center. pp. 165–192. doi:10.1017/CBO9780511762109.006. Heino, J., and Soininen, J. 2007. Are higher taxa adequate surrogates for species-level assemblage patterns and species richness in stream organisms? doi:10.1016/j.biocon.2007.01.017. Herrel, A., and Meijden, A. van der. 2014. An analysis of the live reptile and amphibian trade in the USA compared to the global trade in endangered species. Herpetol. J. 24(8): 103–110. Available from http://www.ingentaconnect.com/content/bhs/thj/2014/00000024/00000002/art00005 [accessed 10 June 2015]. Hilborn, R. 1985. Fleet dynamics and individual variation: why some people catch more fish than others. Can. J. Fish. Aquat. Sci. 42(1): 2–13. Available from http://www.nrcresearchpress.com/doi/abs/10.1139/f85-001 [accessed 9 March 2016]. Hilton-Taylor, C., Pollock, C., and Chanson, J. 2009. State of the world’s species. Threat. Species. Available from https://books.google.ca/books?hl=en&lr=&id=hgErHErt6-gC&oi=fnd&pg=PA15&dq=State+of+the+world’s+species.&ots=ukvqStj6Gm&sig=ivFrG378zvf9xtjH7N5_f6kOLnU [accessed 15 May 2017]. Hong Kong Census and Statistics Department. 2015. Hong Kong Merchandise Trade Statistics Domestic Exports and Re-exports. Hosch, G. 2016. Comparative Analysis of Unilateral and Multilateral Approaches Environment. Geneva.   148 Hughes, T.P., Bellwood, D.R., Folke, C., Steneck, R.S., and Wilson, J. 2005. New paradigms for supporting the resilience of marine ecosystems. Trends Ecol. Evol. 20(7): 380–386. doi:10.1016/j.tree.2005.03.022. Hutchings, J.A., Myers, R.A., García, V.B., Lucifora, L.O., Kuparinen, A., García, V.Β., and Kuparinen4, A. 2012. Life-history correlates of extinction risk and recovery potential. Source Ecol. Appl. 22(224): 1061–1067. Available from http://www.jstor.org [accessed 9 April 2017]. Hutton, J. 2013. Who knows best? Controversy over unilateral stricter domestic measures. In Endangered Species Threatened Convention: The Past, Present and Future of CITES, the Convention on International Trade in Endangered Species of Wild Fauna and Flora. Edited by J.M. Hutton and B. Dickson. Routledge. Available from https://books.google.ca/books?hl=en&lr=&id=2Xr7AQAAQBAJ&oi=fnd&pg=PA57&ots=bnl1uBOdye&sig=ijpt9k5BoJeUJQMV7qBWWuvAfTI [accessed 27 April 2016]. Hutton, J., and Webb, G. 2002. Legal Trade Snaps Back: Using the experience of crocodilians to draw lessons on regulation of the wildlife trade. In Crocodriles. Proceedings of the 16th Working Meeting of the Crocodile specialist Group of the Species Survival Commission of IUCN. Available from http://www.iucncsg.org/365_docs/attachments/protarea/16th-aa4f3b77.pdf#page=22 [accessed 26 April 2017]. ICES. 2010. Report of the Planning Group on Recreational Fisheries (PGRFS). (June): 162. Iles, A. 2007. Making the seafood industry more sustainable: creating production chain transparency and accountability. J. Clean. Prod. 15(6): 577–589. doi:10.1016/j.jclepro.2006.06.001.   149 International Labour Organization and Asian Research Centre for Migration. 2013. Employment practices and working conditions in Thailand’s fishing sector. Jacquet, J., Zeller, D., and Pauly, D. 2010. Counting fish: a typology for fisheries catch data. J. Integr. Environ. Sci. 7(2): 135–144. doi:10.1080/19438151003716498. Johannes, R., Freeman, M., and Hamilton, R. 2000. Ignore fishers’ knowledge and miss the boat. Fish Fish. 1(3): 257–271. Available from http://onlinelibrary.wiley.com/doi/10.1111/j.1467-2979.2000.00019.x/full [accessed 14 November 2015]. Jones, J.P.G., Andriamarovololona, M.M., Hockley, N., Gibbons, J.M., and Milner-Gulland, E.J. 2008. Testing the use of interviews as a tool for monitoring trends in the harvesting of wild species. J. Appl. Ecol. 45(4): 1205–1212. doi:10.1111/j.1365-2664.2008.01487.x. Jones, P.J.S., Qiu, W., and De Santo, E.M. 2013. Governing marine protected areas: Social–ecological resilience through institutional diversity. Mar. Policy 41: 5–13. doi:10.1016/j.marpol.2012.12.026. Kallimanis, A.S., Mazaris, A.D., Tsakanikas, D., Dimopoulos, P., Pantis, J.D., and Sgardelis, S.P. 2012. Efficient biodiversity monitoring: Which taxonomic level to study? Ecol. Indic. 15(1): 100–104. doi:10.1016/j.ecolind.2011.09.024. Kaufmann, D., Kraay, A., and Mastruzzi, M. 2009. Governance Matters VIII: Aggregate and Individual Governance Indicators, 1996-2008. Available from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1424591 [accessed 13 March 2017]. Kremen, C., Cameron, A., Moilanen, A., Phillips, S.J., Thomas, C.D., Beentje, H., Dransfield, J., Fisher, B.L., Glaw, F., Good, T.C., Harper, G.J., Hijmans, R.J., Lees, D.C., Louis, E., Nussbaum, R.A., Raxworthy, C.J., Razafimpahanana, A., Schatz, G.E., Vences, M., Vieites, D.R., Wright, P.C., and Zjhra, M.L. 2008. Aligning Conservation Priorities Across Taxa in   150 Madagascar with High-Resolution Planning Tools. Science (80-. ). 320(5873): 222–226. doi:10.1126/science.1155193. Laksanawimol, P., Petpiroon, S., and Damrongphol, P. 2013. Trade of Seahorses, Hippocampus Spp. (Actinopterygii: Syngnathiformes: Syngnathidae), on the East Coast of the Gulf of Thailand. Acta Ichthyol. Piscat. 43(3): 229–235. Acta Ichthyologica et Piscatoria. doi:10.3750/AIP2013.43.3.07. Lam, J.T.L., Koldewey, H.J., Yasué, M., and Vincent, A.C.J. 2014. Comparing interview and trade data in assessing changes in the seahorse Hippocampus spp. trade following CITES listing. Oryx (November 2002): 1–11. doi:10.1017/S0030605314000246. Lange, D. 1998. Europe’s medicinal and aromatic plants : their use, trade and conservation. Traffic International. Available from https://portals.iucn.org/library/node/7430 [accessed 17 March 2017]. Lascelles, B., Notarbartolo Di Sciara, G., Agardy, T., Cuttelod, A., Eckert, S., Glowka, L., Hoyt, E., Llewellyn, F., Louzao, M., Ridoux, V., and Tetley, M.J. 2014. Migratory marine species: Their status, threats and conservation management needs. Aquat. Conserv. Mar. Freshw. Ecosyst. 24(S2): 111–127. doi:10.1002/aqc.2512. Lawson, J., Foster, S., and Lim, A. 2015. Novel life‐history data for threatened seahorses provide insight into fishery effects. J. Fish Biol. 86(1): 1–15. Available from http://onlinelibrary.wiley.com/doi/10.1111/jfb.12527/pdf [accessed 16 September 2015]. Lawson, J.M., Foster, S.J., and Vincent, A.C.J. 2016. When a little equals a lot – a global assessment of seahorses in bycatch. Fish Fish.   151 Lawson, J.M., Foster, S.J., and Vincent, A.C.J. 2017. Low bycatch rates add up to big numbers for a genus of small fishes. Fisheries 42(1): 19–33. Taylor & Francis. doi:10.1080/03632415.2017.1259944. Lee, D., and Smith, K. 2010. Testudostan: Our post-cold war global exploitation of a noble tortoise. Bull. Chicago Herpetol. Soc. 45: 1–9. Available from http://www.turtlesurvival.org/storage/documents/newsletter_docs/testudostan.pdf [accessed 11 April 2016]. Lee, R.J., Gorog, A.J., Dwiyahreni, A., Siwu, S., Riley, J., Alexander, H., Paoli, G.D., and Ramono, W. 2005. Wildlife trade and implications for law enforcement in Indonesia: a case study from North Sulawesi. Biol. Conserv. 123(4): 477–488. doi:10.1016/j.biocon.2005.01.009. Lenzen, M., Moran, D., Kanemoto, K., Foran, B., Lobefaro, L., and Geschke, A. 2012a. International trade drives biodiversity threats in developing nations. Nature 486: 109–112. doi:10.1038/nature11145. Lenzen, M., Moran, D., Kanemoto, K., Foran, B., Lobefaro, L., and Geschke, A. 2012b. International trade drives biodiversity threats in developing nations. Nature 486(7401): 109–12. Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved. doi:10.1038/nature11145. Leroy, A., Galletti, F., and Chaboud, C. 2016. The EU restrictive trade measures against IUU fishing. Mar. Policy 64: 82–90. doi:10.1016/j.marpol.2015.10.013. Lewison, R., Freeman, S., and Crowder, L. 2004. Quantifying the effects of fisheries on threatened species: the impact of pelagic longlines on loggerhead and leatherback sea   152 turtles. Ecol. Lett. Available from http://onlinelibrary.wiley.com/doi/10.1111/j.1461-0248.2004.00573.x/full [accessed 15 May 2017]. Li, Y., and Li, D. 1998. The dynamics of trade in live wildlife across the Guangxi border between China and Vietnam during 1993–1996 and its control strategies. Biodivers. Conserv. 7(7): 895–914. Kluwer Academic Publishers. doi:10.1023/A:1008873119651. Linneman, H. 1966. An Econometric Study of International Trade Flows. Amsterdam: North-Holland Publishing Company. Available from http://library.wur.nl/WebQuery/clc/457015 [accessed 14 May 2017]. Lobo, A.S., Balmford, A., Arthur, R., and Manica, A. 2010. Commercializing bycatch can push a fishery beyond economic extinction. Conserv. Lett. 3(4): 277–285. doi:10.1111/j.1755-263X.2010.00117.x. Lobo, R., and Jacques, P.J. 2017. SOFIA’S choices: Discourses, values, and norms of the World Ocean Regime. Mar. Policy 78: 26–33. doi:10.1016/j.marpol.2016.12.023. Lozano-Montes, H.M., Pitcher, T.J., and Haggan, N. 2008. Shifting environmental and cognitive baselines in the upper Gulf of California. Front. Ecol. Environ. 6(2): 75–80. doi:10.1890/070056. Lurié, A., and Kalinina, M. 2015. Protecting Animals in International Trade: A Study of the Recent Successes at the WTO and in Free Trade Agreements. Am. Univ. Int. Law Rev. 30(3): 431. American University. Available from https://www.questia.com/library/journal/1P3-3704630821/protecting-animals-in-international-trade-a-study [accessed 21 April 2016].   153 Lyons, J.E., Runge, M.C., Laskowski, H.P., and Kendall, W.L. 2008. Monitoring in the context of structured decision-making and adaptive management. J. Wildl. Manage. 72(8): 1683–1692. doi:10.2193/2008-141. MacMillan, D.C., and Han, J. 2011. Cetacean By-Catch in the Korean Peninsula—by Chance or by Design? Hum. Ecol. 39(6): 757–768. Springer US. doi:10.1007/s10745-011-9429-4. Martin-Smith, K., Samoilys, M., and Meeuwig, J. 2004. Collaborative development of management options for an artisanal fishery for seahorses in the central Philippines. Ocean Coast. Available from http://www.sciencedirect.com/science/article/pii/S0964569104000298 [accessed 15 May 2017]. Maxwell, S., Fuller, R., Brooks, T., and Watson, J. 2016. The ravages of guns, nets and bulldozers. Nature 536(7615): 143–145. Available from http://www.ncbi.nlm.nih.gov/pubmed/27510207 [accessed 23 August 2016]. McClenachan, L., Cooper, A.B., and Dulvy, N.K. 2016. Rethinking Trade-Driven Extinction Risk in Marine and Terrestrial Megafauna. Curr. Biol. 26(0): 1–7. Elsevier. doi:10.1016/j.cub.2016.05.026. McDonald, J. 2014. Handbook of Biological Statistics. In 3rd ed. Sparky House Publishing, Baltimore, Maryland. Available from http://www.biostathandbook.com/kruskalwallis.html [accessed 22 April 2017]. McShane, P.E. 1996. Patch dynamics and effects of exploitation on abalone (Haliotis iris) populations. Fish. Res. 25(2): 191–199. doi:10.1016/0165-7836(95)00416-5. Mejía, E., Buitrón, X., Peña-Claros, M., and Grogan, J. 2008. Big-leaf mahogany (Swietenia macrophylla) in peru, Bolivia and Brazil. In International expert workshop on CITES non-  154 detriment findings. Cancun, Mexico. Available from http://www.conabio.gob.mx/institucion/cooperacion_internacional/TallerNDF/Links-Documentos/WG-CS/WG1-Trees/WG1-CS4 Swietenia/WG1-CS4-P.pdf [accessed 27 April 2016]. Milner-Gulland, E.J., and Bennett, E.L. 2003. Wild meat: the bigger picture. Trends Ecol. Evol. 18(7): 351–357. doi:10.1016/S0169-5347(03)00123-X. Di Minin, E., Laitila, J., Montesino-Pouzols, F., Leader-Williams, N., Slotow, R., Goodman, P.S., Conway, A.J., and Moilanen, A. 2015. Identification of policies for a sustainable legal trade in rhinoceros horn based on population projection and socioeconomic models. Conserv. Biol. 29(2): 545–555. doi:10.1111/cobi.12412. Mittermeier, R.A., Goettsch Mittermeier, C., and Robles Gil, P. 1997. Megadiversity : earth’s biologically wealthiest nations. CEMEX. Moore, J.E., Cox, T.M., Lewison, R.L., Read, A.J., Bjorkland, R., McDonald, S.L., Crowder, L.B., Aruna, E., Ayissi, I., Espeut, P., Joynson-Hicks, C., Pilcher, N., Poonian, C.N.S., Solarin, B., and Kiszka, J. 2010. An interview-based approach to assess marine mammal and sea turtle captures in artisanal fisheries. Biol. Conserv. 143(3): 795–805. doi:10.1016/j.biocon.2009.12.023. Moore, J.E., Curtis, K.A., Lewison, R.L., Dillingham, P.W., Cope, J.M., Fordham, S. V., Heppell, S.S., Pardo, S.A., Simpfendorfer, C.A., Tuck, G.N., and Zhou, S. 2013. Evaluating sustainability of fisheries bycatch mortality for marine megafauna: a review of conservation reference points for data-limited populations. Environ. Conserv. 40(4): 329–344. Cambridge University Press. doi:10.1017/S037689291300012X.   155 Moran, D., and Kanemoto, K. 2017. Identifying species threat hotspots from global supply chains. Nat. Ecol. Evol. 1(1): 23. Nature Publishing Group. doi:10.1038/s41559-016-0023. Morwitz, V. 1997. It seems like only yesterday: The nature and consequences of telescoping errors in marketing research. J. Consum. Psychol. 6(1): 1–29. Available from http://www.sciencedirect.com/science/article/pii/S1057740897703772 [accessed 18 October 2015]. Mundy-Taylor, V., Crook, V., Foster, S., Fowler, S., and Sant, G. 2014. CITES Non-detriment Findings Guidance for Shark Species. Available from http://www.cites.org/sites/default/files/common/com/ac/27/E-AC27-Inf-01.pdf [accessed 25 April 2017]. Naidoo, R., Balmford, A., Costanza, R., Fisher, B., Green, R.E., Lehner, B., Malcolm, T.R., and Ricketts, T.H. 2008. Global mapping of ecosystem services and conservation priorities. Proc. Natl. Acad. Sci. U. S. A. 105(28): 9495–500. National Academy of Sciences. doi:10.1073/pnas.0707823105. Nash, S. V. 1993. Sold for a Song: The Trade in Southeast Asian Non-CITES Birds. Natale, F., Borrello, A., and Motova, A. 2015. Analysis of the determinants of international seafood trade using a gravity model. Mar. Policy 60: 98–106. doi:10.1016/j.marpol.2015.05.016. Neis, B., Schneider, D.C., Felt, L., Haedrich, R.L., Fischer, J., and Hutchings, J. a. 1999. Fisheries assessment: what can be learned from interviewing resource users? Can. J. Fish. Aquat. Sci. 56(10): 1949–1963. doi:10.1139/f99-115. Nijman, V. 2010a. An overview of international wildlife trade from Southeast Asia. Biodivers. Conserv. 19: 1101–1114. doi:10.1007/s10531-009-9758-4.   156 Nijman, V. 2010b. An overview of international wildlife trade from Southeast Asia. Biodivers. Conserv. CN  - CITES2010.003 19(4): 1101–1114. Nijman, V., Nekaris, K. a I., Donati, G., Bruford, M., and Fa, J. 2011. Primate conservation: Measuring and mitigating trade in primates. Endanger. Species Res. 13(Smith 1978): 159–161. doi:10.3354/esr00336. Nijman, V., and Shepherd, C.R. 2007. Trade in non-native, CITES-listed, wildlife in Asia, as exemplified by the trade in freshwater turtles and tortoises (Chelonidae) in Thailand. Contrib. to Zool. 76(3): 207–211. Nijman, V., and Shepherd, C.R. 2010. The role of Asia in the global trade in CITES II-listed poison arrow frogs: Hopping from Kazakhstan to Lebanon to Thailand and beyond. Biodivers. Conserv. 19(7): 1963–1970. doi:10.1007/s10531-010-9814-0. Nijman, V., and Shepherd, C.R. 2011a. The role of Thailand in the international trade in CITES-listed live reptiles and amphibians. PLoS One 6(3): 1–7. doi:10.1371/journal.pone.0017825. Nijman, V., and Shepherd, C.R. 2011b. The role of Thailand in the international trade in CITES-listed live reptiles and amphibians. PLoS One 6(3): e17825. Public Library of Science. doi:10.1371/journal.pone.0017825. Nunoo, F.K.E., Boateng, J.O., Ahulu, A.M., Agyekum, K.A., and Sumaila, U.R. 2009. When trash fish is treasure: The case of Ghana in West Africa. Fish. Res. 96: 167–172. doi:10.1016/j.fishres.2008.10.010. O’Donnell, K.P., Molloy, P.P., and Vincent, A.C.J. 2012. Comparing fisher Interviews, Logbooks, and Catch Landings Estimates of Extraction Rates in a Small-Scale Fishery. Coast. Manag. 40(6): 594–611. doi:10.1080/08920753.2012.727734.   157 O’Donnell, K.P., Pajaro, M.G., and Vincent,  a. C.J. 2010. How does the accuracy of fisher knowledge affect seahorse conservation status? Anim. Conserv. 13(6): 526–533. doi:10.1111/j.1469-1795.2010.00377.x. OECD. 2000. Trade Measures in Multilateral Environmental Agreements. OECD Publishing. Available from https://books.google.com/books?id=IprWAgAAQBAJ&pgis=1 [accessed 1 April 2016]. OECD. 2015. OECD Review of Fisheries: Country Statistics, OECD Review of Fisheries: Country Statistics 2014. OECD Publishing, Paris. Available from http://www.oecd-ilibrary.org/agriculture-and-food/oecd-review-of-fisheries-country-statistics-2014_rev_fish_stat_en-2014-en [accessed 18 January 2016]. Oldfield, S. 2003. The trade in wildlife: regulation for conservation. Earthscan Publications Ltd. Orensanz, J.M., Armstrong, J., Armstrong, D., and Hilborn, R. 1998. Crustacean resources are vulnerable to serial depletion – the multifaceted decline of crab and shrimp fisheries in the Greater Gulf of Alaska. Rev. Fish Biol. Fish. 8(2): 117–176. Kluwer Academic Publishers. doi:10.1023/A:1008891412756. Oriental Daily. 2014. 珊瑚炫富海馬壯陽 經港走私內地激增 - 東方日報. Available from http://orientaldaily.on.cc/cnt/news/20140623/00176_033.html [accessed 26 January 2017]. Orme, C.D.L., Davies, R.G., Burgess, M., Eigenbrod, F., Pickup, N., Olson, V.A., Webster, A.J., Ding, T.-S., Rasmussen, P.C., Ridgely, R.S., Stattersfield, A.J., Bennett, P.M., Blackburn, T.M., Gaston, K.J., and Owens, I.P.F. 2005. Global hotspots of species richness are not congruent with endemism or threat. Nature 436(7053): 1016–1019. doi:10.1038/nature03850.   158 Panjarat, S., and Bennett, N. 2012. Responses of fishers to a 25-year seasonal closed measure on the Andaman Coast of Thailand. In Project IMPAACT, Marine Protected Areas Research Group, University of Victoria. Available from http://www.researchgate.net/profile/Nathan_Bennett3/publication/262689448_Responses_of_fishers_to_a_25-year_seasonal_closed_measure_on_the_Andaman_Coast_of_Thailand/links/0f31753a2173616178000000.pdf [accessed 15 November 2015]. Patel, N.G. 2015. Characterization of illegal wildlife trade networks. Available from http://repository.upenn.edu/edissertations/1115 [accessed 6 April 2017]. Patel, N.G., Rorres, C., Joly, D.O., Brownstein, J.S., Boston, R., Levy, M.Z., and Smith, G. 2015. Quantitative methods of identifying the key nodes in the illegal wildlife trade network. Proc. Natl. Acad. Sci. U. S. A. 112(26): 7948–53. doi:10.1073/pnas.1500862112. Patrick, W., Spencer, P., Ormseth, O., Cope, J., Field, J., Kobayashi, D., Gedamke, T., Coretes, E., Bigelow, K., Overholtz, W., Link, J., and Lawson, P. 2009. Use of productivity and susceptibility indices to determine the vulnerability of a stock: with example applications to six US fisheries. In NOAA Tech. Memo. Available from ftp://ftp.pcouncil.org/pub/Briefing Books/BB_CDs/BB_CD_0609/0609/E4a_ATT1_0609.pdf [accessed 25 November 2015]. Pauly, D., Christensen, V., Guénette, S., Pitcher, T.J., Sumaila, U.R., Walters, C.J., Watson, R., and Zeller, D. 2002. Towards sustainability in world fisheries. Nature 418(6898): 689–695. Nature Publishing Group. doi:10.1038/nature01017. Pernetta, A. 2009. Monitoring the trade: using the CITES database to examine the global trade in live monitor lizards (Varanus spp.). Biawak 3(2): 37–45. Available from   159 http://www.tsusinvasives.org/dotAsset/24542536-c835-4750-8a75-5bfc2fc5e9eb.pdf [accessed 26 January 2017]. Perry, A.L., Lunn, K.E., and Vincent, A.C.J. 2010. Fisheries, large-scale trade, and conservation of seahorses in Malaysia and Thailand. Aquat. Conserv. Mar. Freshw. Ecosyst. 20(4): 464–475. doi:10.1002/aqc.1112. Phelps, J., Webb, E.L., Bickford, D., Nijman, V., and Sodhi, N.S. 2010. Boosting CITES. Science (80-. ). 330(6012): 1752–3. doi:10.1126/science.1195558. Phoonsawat, R., Boonjorn, N., Sinanuan, T., Chamason, O., Kulanujaree, N., Jaiyen, T., Thodsapornpithakkul, S., Kjandara, T., Sumontra, M., Boonsuk, S., and Nootmorn, P. 2015. Report on Thailand’s actions addressing problems of Hippocampus spp. Available from https://cites.org/sites/default/files/eng/com/sc/66/E-SC66-31-01_A3.pdf [accessed 25 January 2016]. Poole, C.M., and Shepherd, C.R. 2016. Shades of grey: the legal trade in CITES-listed birds in Singapore, notably the globally threatened African grey parrot Psittacus erithacus. Oryx: 1–7. doi:10.1017/S0030605314000234. Pramod, G., Nakamura, K., Pitcher, T., and Delagran, L. 2014. Estimates of illegal and unreported fish in seafood imports to the USA. Mar. Policy. Available from http://www.sciencedirect.com/science/article/pii/S0308597X14000918 [accessed 4 May 2017]. Pusceddu, A., Bianchelli, S., Martín, J., Puig, P., Palanques, A., Masqué, P., and Danovaro, R. 2014. Chronic and intensive bottom trawling impairs deep-sea biodiversity and ecosystem functioning. Proc. Natl. Acad. Sci. U. S. A. 111(24): 8861–6. National Academy of Sciences. doi:10.1073/pnas.1405454111.   160 Rao, R., Glickman, M., and Glynn, R. 2004. Use of covariates and survey wave to adjust for nonresponse. Biometrical J. 46(5): 579–588. Available from http://onlinelibrary.wiley.com/doi/10.1002/bimj.200310049/abstract [accessed 18 October 2015]. Raymakers, C., and Hoover, C. 2002. Acipenseriformes: CITES implementation from Range States to consumer countries. J. Appl. Ichthyol. 18(4–6): 629–638. Blackwell Science Ltd. doi:10.1046/j.1439-0426.2002.00398.x. Recharte Uscamaita, M., and Bodmer, R. 2010a. Recovery of the Endangered giant otter Pteronura brasiliensis on the Yavarí-Mirín and Yavarí Rivers: a success story for CITES. Oryx 44(1): 83. doi:10.1017/S0030605309990196. Recharte Uscamaita, M., and Bodmer, R. 2010b. Recovery of the Endangered giant otter Pteronura brasiliensis on the Yavarí-Mirín and Yavarí Rivers: a success story for CITES. Oryx 44(1): 83. doi:10.1017/S0030605309990196. Reeve, R. 2006. Wildlife trade, sanctions and compliance: lessons from the CITES regime. Int. Aff. CN  - CITES2006.002 82(5): 881–897. Regan, H., and Ben-Haim, Y. 2005. Robust decision-making under severe uncertainty for conservation management. Ecol. Appl. 15(4): 1471–1477. Available from http://www.esajournals.org/doi/abs/10.1890/03-5419 [accessed 24 October 2015]. Rice, J. 2007. International Trade and Uneven Ecological Unequal Exchange: Utilization of Environmental Space in the World System. Soc. Forces 85(3): 1369–1392. doi:10.1353/sof.2007.0054.   161 Rist, J., Milner-Gulland, E.J., Cowlishaw, G., and Rowcliffe, M. 2010. Hunter reporting of catch per unit effort as a monitoring tool in a bushmeat-harvesting system. Conserv. Biol. 24(2): 489–99. doi:10.1111/j.1523-1739.2010.01470.x. Rivalan, P., Delmas, V., Angulo, E., Bull, L.S., Hall, R.J., Courchamp, F., Rosser, A.M., and Leader-Williams, N. 2007. Can bans stimulate wildlife trade? Nature 447(7144): 529–530. Nature Publishing Group. doi:10.1038/447529a. Robinson, J.E., Griffiths, R. a., St. John, F.A.V., and Roberts, D.L. 2015. Dynamics of the global trade in live reptiles: Shifting trends in production and consequences for sustainability. Biol. Conserv. 184: 42–50. Elsevier Ltd. doi:10.1016/j.biocon.2014.12.019. Roe, D., Mulliken, T., Milledge, S., Mremi, J., Mosha, S., and Grieg-gran, M. 2002. Making a killing or making a living ? Wildlife trade , trade controls and rural livelihoods. Biodivers. Livelihoods 59(6): 109. Roth, W., and Mehta, J. 2002. The Rashomon effect combining positivist and interpretivist approaches in the analysis of contested events. Sociol. Methods Res. 31(2): 131–173. Available from http://smr.sagepub.com/content/31/2/131.short [accessed 15 November 2015]. Round, P.D. 1990. Bangkok bird club survey of the bird and mammal trade in the Bangkok weekend market. Nat. Hist. Bull. Siam Soc. 38: 1–3. Available from http://www.siamese-heritage.org/nhbsspdf/vol031-040/NHBSS_038_1e_Round_BangkokBirdClubSu.pdf [accessed 16 March 2017]. Sadovy, Y. 2010. Humphead wrasse and illegal, unreported and unregulated fishing. SPC Live Reef Fish Inf. Bull. 19. Available from   162 http://www.spc.int/DigitalLibrary/Doc/FAME/InfoBull/LRF/19/LRF19_19_Sadovy.pdf [accessed 27 April 2016]. Sadovy, Y., Punt, A.E., Cheung, W., Vasconcellos, M., Suharti, S., and Mapstone, B.D. 2007. Stock assessment approach for the Napoleon fish, Cheilinus undulatus, in Indonesia. A tool for quota-setting for poor-data fisheries under CITES Appendix II Non-Detriment Finding requirements. FAO Fish. Circ. FAO. Available from http://agris.fao.org/agris-search/search.do?recordID=XF2008434187 [accessed 27 April 2016]. Salin, K. 2005. Fisheries and trade of seahorses, Hippocampus spp., in southern India. Fish. Manag. Ecol. 12(4): 269–273. Available from http://onlinelibrary.wiley.com/doi/10.1111/j.1365-2400.2005.00450.x/full [accessed 15 November 2015]. Sand, P.H. 2013. Enforcing CITES: The Rise and Fall of Trade Sanctions. Rev. Eur. Comp. Int. Environ. Law 22(3): 251–263. doi:10.1111/reel.12037. Scales, H., Balmford, A., and Manica, A. 2007. Monitoring the live reef food fish trade: Lessons learned from local and global perspectives. SPC Live Reef Fish Inf. Bull. 17: 36–44. Available from http://www.ecocean.fr/wp-content/uploads/Ecocean-Elevage-Raisonne-Article-Lecaillon-2007-en.pdf#page=36 [accessed 9 June 2015]. Schlaepfer, M. a., Hoover, C., and Dodd, C.K. 2005. Challenges in Evaluating the Impact of the Trade in Amphibians and Reptiles on Wild Populations. Bioscience 55(3): 256. doi:10.1641/0006-3568(2005)055[0256:CIETIO]2.0.CO;2. Schmidt, J., Kellie, K., and Chapin, F. 2015. Detecting, estimating, and correcting for biases in harvest data. J. Wildl. Manage. 79(7): 1152–1162. Available from http://onlinelibrary.wiley.com/doi/10.1002/jwmg.928/full [accessed 14 October 2015].   163 Shepherd, C.R., and Nijman, V. 2007. The trade in bear parts from Myanmar: an illustration of the ineffectiveness of enforcement of international wildlife trade regulations. Biodivers. Conserv. 17(1): 35–42. doi:10.1007/s10531-007-9228-9. Shepperson, J., Murray, L., and Cook, S. 2014. Methodological considerations when using local knowledge to infer spatial patterns of resource exploitation in an Irish Sea fishery. Biol. Conserv. 180: 214–223. Available from http://www.sciencedirect.com/science/article/pii/S0006320714003942 [accessed 15 November 2015]. Smith, K., Behrens, M., and Schloegel, L. 2009a. Reducing the risks of the wildlife trade. Science (80-. ). Available from http://www.planta.cn/forum/files_planta/science_2009_smith_594_5_157.pdf [accessed 10 February 2016]. Smith, K.F., Behrens, M., Schloegel, L.M., Marano, N., Burgiel, S., and Daszak, P. 2009b. Ecology. Reducing the risks of the wildlife trade. Science (80-. ). 324(May): 594–595. doi:10.1126/science.1174460. Smith, K.F., Behrens, M., Schloegel, L.M., Marano, N., Burgiel, S., and Daszak, P. 2009c. Reducing the risks of the wildlife trade. Science (80-. ). 324(5927). Available from http://science.sciencemag.org/content/324/5927/594.full [accessed 11 April 2017]. Smith, M., and Benítez-Díaz, H. 2011. Assessing the impacts of international trade on CITES-listed species: current practices and opportunities for scientific research. Biol.  …. Available from http://www.sciencedirect.com/science/article/pii/S0006320710004611 [accessed 14 September 2015].   164 Smith, M.J., Benítez-Díaz, H., Clemente-Muñoz, M.Á., Donaldson, J., Hutton, J.M., Noel McGough, H., Medellin, R. a., Morgan, D.H.W., O’Criodain, C., and Oldfield, T.E.E. 2011. Assessing the impacts of international trade on CITES-listed species: Current practices and opportunities for scientific research. Biol. Conserv. 144(1): 82–91. Elsevier Ltd. doi:10.1016/j.biocon.2010.10.018. Song, N. Van. 2008. Wildlife Trading in Vietnam Situation, Causes, and Solutions. J. Environ. Dev. 17(2): 145–165. doi:10.1177/1070496508316220. Sorice, M., Oh, C., and Gartner, T. 2013. Increasing participation in incentive programs for biodiversity conservation. Ecol. Appl. 23(5): 1146–1155. Available from http://www.esajournals.org/doi/abs/10.1890/12-1878.1 [accessed 20 January 2016]. Srikosamatara, S., Siripholdej, B., and Suteethorn, V. 1992. Wildlife trade in Lao PDR and between Lao PDR and Thailand. Nat. Hist. Bull. Siam … 40(1): 47. Available from https://scholar.google.ca/scholar?q=Wildlife+trade+in+Lao+PDR+and+between+Lao+PDR+and+Thailand&btnG=&hl=en&as_sdt=0%252C5#0 [accessed 25 January 2016]. Sterner, T. 2003. Policy instruments for environmental and natural resource management. Resources for the Future. Available from https://books.google.ca/books/about/Policy_Instruments_for_Environmental_and.html?id=Cwnsdgk3PY4C&redir_esc=y [accessed 15 May 2017]. Sudman, S., and Bradburn, N. 1974. Response effects in surveys: a review and synthesis. In Chicago: Aldine Pub. Co . Sullivan, M.G. 2003. Exaggeration of Walleye Catches by Alberta Anglers. North Am. J. Fish. Manag. 23(2): 573–580.  Taylor & Francis Group . doi:10.1577/1548-8675(2003)023<0573:EOWCBA>2.0.CO;2.   165 Swartz, W., Sumaila, U.., Watson, R., and and Pauly, D. 2010. Sourcing Seafood for the Three Major Markets:The EU,Japan and the USA. Mar. Policy 34: 1366–1373. Available from https://www.researchgate.net/profile/Rashid_Sumaila/publication/222650899_Sourcing_seafood_for_the_three_major_markets_The_EU_Japan_and_the_USA/links/09e41505b58ce6a6ea000000.pdf [accessed 22 March 2017]. Tarrant, M.A., Manfredo, M.J., Bayley, P.B., and Hess, R. 1993. Effects of recall bias and nonresponse bias on self-report estimates of angling participation. North Am. J. Fish. Manag. 13(2): 217–222. Taylor & Francis Group. doi:10.1577/1548-8675(1993)013<0217:EORBAN>2.3.CO;2. Taylor, R.B., Morrison, M. a., and Shears, N.T. 2011. Establishing baselines for recovery in a marine reserve (Poor Knights Islands, New Zealand) using local ecological knowledge. Biol. Conserv. 144(12): 3038–3046. Elsevier Ltd. doi:10.1016/j.biocon.2011.09.009. Teh, L., Zeller, D., and Pauly, D. 2015. Preliminary reconstruction of Thailand’s fisheries catches: 1950-2010. Vancouver, BC. Teh, L.C.L., and Sumaila, U.R. 2013. Contribution of marine fisheries to worldwide employment. Fish Fish. 14(1): 77–88. Blackwell Publishing Ltd. doi:10.1111/j.1467-2979.2011.00450.x. Tesfamichael, D., Pitcher, T.J., and Pauly, D. 2014. Assessing changes in fisheries using fishers’ knowledge to generate long time series of catch rates: A case study from the red sea. Ecol. Soc. 19(1): 18. The Resilience Alliance. The World Bank. 2016. GDP per capita. Available from http://data.worldbank.org/indicator/NY.GDP.PCAP.CD [accessed 16 May 2016].   166 Thorbjarnarson, J. 1999. Crocodile Tears and Skins: International Trade, Economic Constraints, and Limits to the Sustainable Use of Crocodilians. Conserv. Biol. 13(3): 465–470. doi:10.1046/j.1523-1739.1999.00011.x. Thornton, T., and Scheer, A. 2012. Collaborative engagement of local and traditional knowledge and science in marine environments: a review. Ecol. Soc. 17(3): 8. Available from http://dlc.dlib.indiana.edu/dlc/handle/10535/8574 [accessed 22 September 2015]. Tobias, T.N. 2010. Living proof: the essential data collection guide for indigenous use and occupancy map surveys. Ecotrust Canada and Union of British Columbia Indian Chiefs (UBCIC). Tolotti, M.T., Filmalter, J.D., Bach, P., Travassos, P., Seret, B., and Dagorn, L. 2015. Banning is not enough: The complexities of oceanic shark management by tuna regional fisheries management organizations. In Global Ecology and Conservation. doi:10.1016/j.gecco.2015.05.003. Toms, J.D., and Lesperance, M.L. 2003. Picewise regression: a tool for identifying ecological thresholds. Ecology 84(8): 2034–2041. Ecological Society of America. doi:10.1890/02-0472. TRAFFIC. 2008. What’s driving the wildlife trade? A review of expert opinion on economics and social drivers of the wildlife trade and trade control efforts in Cambodia, Indonesia, Lao PDR and Vietnam. In East Asia and Pacific Region Sustainabl Development Discussion Papers. East Asia and Pacific Region Sustainable Development Department, World Bank, Washington, DC. Available from file:///Users/cot/Downloads/traffic_pub_gen24.pdf [accessed 11 June 2015]. UNEP-WCMC. 2007. Indonesia : A review of trade in CITES- listed species.   167 UNEP-WCMC. 2013a. CITES Trade Database. Available from http://trade.cites.org/ [accessed 27 April 2016]. UNEP-WCMC. 2013b. A guide to using the CITES Trade Database. Available from https://trade.cites.org/cites_trade_guidelines/en-CITES_Trade_Database_Guide.pdf [accessed 7 March 2017]. UNEP-WCMC. 2015. The Checklist of CITES Species Website. Available from http://checklist.cites.org [accessed 21 January 2017]. Usher, P. 2000. Traditional ecological knowledge in environmental assessment and management. Arctic: 183–193. Available from http://www.jstor.org/stable/40512207 [accessed 14 November 2015]. Varkey, D. a., Ainsworth, C.H., Pitcher, T.J., Goram, Y., and Sumaila, R. 2010. Illegal, unreported and unregulated fisheries catch in Raja Ampat Regency, Eastern Indonesia. Mar. Policy 34(2): 228–236. Elsevier. doi:10.1016/j.marpol.2009.06.009. Vaske, J., Huan, T.-C., and Beaman, J. 2003. The Use of Multiples in Anglers’ Recall of Participation and Harvest Estimates: Some Results and Implications. Leis. Sci. 25(4): 399–409. Taylor & Francis. doi:10.1080/714044498. Velázquez Gomar, J.O., and Stringer, L.C. 2011. Moving towards sustainability? An analysis of CITES’ conservation policies. Environ. Policy Gov. 21(4): 240–258. doi:10.1002/eet.577. Vincent, A. 1996. The international trade in seahorses. TRAFFIC Int. Available from http://orton.catie.ac.cr/cgi-bin/wxis.exe/?IsisScript=BIRENA.xis&method=post&formato=2&cantidad=1&expresion=mfn=000803 [accessed 10 December 2013].   168 Vincent, A.C.J., and Foster, S.J. 2017. Setting Precedent in Export Regulations for Marine Fishes with Seahorses. Fisheries 42(1): 40–43. Taylor & Francis. doi:10.1080/03632415.2017.1273658. Vincent, A.C.J., Sadovy de Mitcheson, Y.J., Fowler, S.L., and Lieberman, S. 2013. The role of CITES in the conservation of marine fishes subject to international trade. Fish Fish.: n/a-n/a. doi:10.1111/faf.12035. Vincent,  a. C.J., Foster, S.J., and Koldewey, H.J. 2011. Conservation and management of seahorses and other Syngnathidae. J. Fish Biol. 78: 1681–1724. doi:10.1111/j.1095-8649.2011.03003.x. Walsh, P.D., and White, L.J.T. 1999. What It Will Take to Monitor Forest Elephant Populations. Conserv. Biol. 13(5): 1194–1202. Blackwell Science Inc. doi:10.1046/j.1523-1739.1999.98148.x. Walters, C.J., and Holling, C.S. 1990. Large-Scale Management Experiments and Learning by Doing. Ecology 71(6): 2060–2068. Ecological Society of America. doi:10.2307/1938620. Watkins, K., and Fowler, P. 2002. Rigged Rules and Double Standards: Trade Globalisation, and the Fight Against Poverty. Oxfam Int. (Oxfam Campaign Reports): 276. doi:9780855985257. Watson, R.A., Nichols, R., Lam, V.W.Y., and Sumaila, U.R. 2017. Global seafood trade flows and developing economies: Insights from linking trade and production. Mar. Policy 82: 41–49. doi:10.1016/j.marpol.2017.04.017. Watts, D.J., and Strogatz, S.H. 1998. Collective dynamics of “small- world” networks. Nature 393(6684): 440–442. Available from   169 http://www.nature.com/nature/journal/v393/n6684/abs/393440a0.html [accessed 7 April 2017]. Wiersema, A. 2013. Uncertainty and Markets for Endangered Species under CITES. Rev. Eur. Comp. Int. Environ. Law 22(3): 239–250. doi:10.1111/reel.12047. Willman, R., Kelleher, K., Arnason, R., and Franz, N. 2012. The sunken billions: the economic justification for fisheries reform. Mar. Resour. Econ. 27(2): 193–195. Available from http://agris.fao.org/agris-search/search.do?recordID=XF2009438185 [accessed 25 January 2016]. Wood, E., Malsch, K., and Miller, J. 2012. International trade in hard corals: review of management, sustainability and trends. 12th Int. Coral Reef Symp. (January): 9–13. Wu, J. 2016. Shark fin and mobulid ray gill plate trade in mainland China, Hong Kong and Taiwan. Taipei. Yasué, M., Nellas, A., Panes, H., and Vincent, A. 2015. Monitoring landed seahorse catch in a changing policy environment. Endanger. Species Res. 27: 95–111. Available from http://www.int-res.com/abstracts/esr/v27/n2/p95-111/ [accessed 26 January 2016]. Young, O.R. 2011. Effectiveness of international environmental regimes: existing knowledge, cutting-edge themes, and research strategies. Proc. Natl. Acad. Sci. U. S. A. 108(50): 19853–60. National Academy of Sciences. doi:10.1073/pnas.1111690108. Zarauz, L., Ruiz, J., Urtizberea, A., Andonegi, E., Mugerza, E., and Artetxe, I. 2015. Comparing different survey methods to estimate European sea bass recreational catches in the Basque Country. ICES J. Mar. Sci. 72(4): 1181–1191. doi:10.1093/icesjms/fsv054.   170 Zhang, L., Hua, N., and Sun, S. 2008. Wildlife trade, consumption and conservation awareness in southwest China. Biodivers. Conserv. 17(6): 1493–1516. doi:10.1007/s10531-008-9358-8.    171 Appendices  Appendix A  Supporting material for Chapter 2 A.1 Corroboration with governmental trawl survey data We corroborated our fisher-reported annual catch rates with catch rates from a three-year dataset of government research trawls from the Thailand Department of Fisheries (DoF). Since the DoF trawl surveys were conducted by otter trawl, we only compared this data to our interviews from otter trawl fishers. Trawl surveys were executed by DoF in 2010, 2012 and 2013 at pre-determined sampling locations throughout the Andaman Sea and Gulf of Thailand (Fig. S1). These locations were originally intended for the purpose of sampling commercially exploited fish species, but recording the presence or absence of seahorses for each research trawl began in 2010 (Table S1). Each location was sampled four times per year using a 23.5 m otter-board trawl research vessel, with trawl speed set at 2.5 nautical miles/hour. Research trawls took place within each location for one hour. In order to compare this effort to our reported trawl fisher effort (mean = 4 hours/haul), we multiplied the DoF catch rate by 4 (Table S2).  We analyzed the DoF trawl data based on two methods to weight catch rates by coast (Table S2). With our first method, we divided the total number of seahorses by the total number of hauls, despite unequal survey effort for each coast. With this analysis, we assumed that (1) the larger survey effort in the Gulf of Thailand was representative of a larger proportion of fishers on this coast and as in our data (33 trawlers from Gulf Coast and 8 from Andaman Sea), and (2) fishers did not have an opportunity to fish on either coast. For our second analysis we weighted catch rates by coast equally to determine a national average. This analysis assumes that fishers have an   172 equal chance to fish on both coasts, and there are two separate populations of seahorses, one on the Andaman Sea and one on the Gulf Coast.  Table A.1 Department of Fisheries Governmental trawls from 2010, 2012 and 2013. Year Total Grids Searched # Grids with SH presence # hauls total / year  Total # of SH Mean SH / haul  Gulf of Thailand  2010 63 16 252 73 0.29 2012 63 4 252 21 0.08 2013 63 2 252 2 0.007 Andaman   2010 22 7 88 14 0.16 2012 22 20 88 519 5.9 2013 22 4 88 5 0.06   173 Table A.2 Mean and median of annual catch rates per trawler based on Department of Fisheries’ surveys over a three year period. The annual rate was based on the catch rates for 4-hours haul.   SH / haul SH/4-hrs haul Hauls / Day Trip minimum Trip / month Mo / year Annual catch estimate 1) unequal weighting  by coast DoF Mean 0.6 2.4 4 20 1.3 11.1 2,771 DoF Median 0 0 4 20 1.3 11.1 0 2) equal weighting by coast DoF Mean 1.1 4.4 4 20 1.3 11.1 5,079 DoF Median 0 0 4 20 1.3 11.1 0   174 Table A.3 Comparison of annual catch rates from Thai Department of Fisheries research trawls and fisher interviews in this study. Both mean and median annual catch per trawler are presented.   DoF research trawls Trawl fisher interviews (Otter trawlers)   Reporting time period  DoF Trawls (equal) DoF Trawls (unequal) All interviews (n=34) Per-haul (n=12) Per-day (n=8) Per-trip (n=17) Per-month (n=5) Per-year (n=2) Mean 5,079 2,771 4,977 13,577 1,682 1,269 991 1 Median 0 0 1,695 3,984 1,731 1,050 120 1   175 Table A.4 Comparison of annual seahorse catch rates for Trat province from fisher interviews (this study) and port-sampling (Laksanawimol et al. 2013). Trawler Gillnet  Reported as Annual catch  n Reported as Annual catch  n This study Per-trip 246 1 Per-day 16 10     Per-month 365 4 Laksanawimol et al. 2013 Sampled by fishers and recorded every month 219 20 Sampled by fishers and recorded every month 42 15   176 Table A.4 Summary table shows the numbers of seahorse catch rates reported in previously published studies by their reporting time periods. Location Gear Time period reported in paper Haul Day Week Trip Month Year Source Florida Trawls Day 1 1 -- -- -- -- Baum et al. 2003 Central /  South America Various Annual 1 -- -- 1 3 4 Baum et al. 2005 Malaysia Trawls Annual 1 1 1 -- 1 -- Choo et al. 2005 Vietnam Trawls Day & annual -- 1 -- -- -- -- Giles et al. 2006 Thailand Various Annual -- -- -- 1 1 -- Laksanawimol et al. 2013 Global Various Day & annual 2 12 6 1 26 25 Lawson et al., 2017* Vietnam Trawls Day & annual -- 1 -- -- 1 -- Meeuwig et al. 2006 India Seines &  Trawls Monthly & annual -- 1 -- -- 1 -- Murugan et al. 2001 Malaysia, Thailand Trawls Annual -- -- 1 1 1 1 Perry et al. 2010 Brazil By hand Annual -- 1 -- -- -- -- Rosa et al. 2006 India Trawls Annual -- 1 -- -- 1 -- Salin et al. 2005   Total 5 19 8 4 35 30  * Only unpublished reports of seahorse catch surveys were used from this paper in order to avoid double counting.    177 Table A.5 Previously published studies reporting seahorse catch per haul but annual estimates based on another time period. We calculated annual estimates from reported haul values to compare with those published based on other periods. Reported seahorses per haul Annual estimates from reported haul values Reported seahorse catch/ year Location Gear Source 0-16 (Port sampling) 59,520-1,071,360 72,000 Hernando Beach, Florida Trawl (Baum et al. 2003a) 20-100 1,000,000-6,000,000 30,000-72,000 Guayaquil, Ecuador Trawl (Baum and Vincent 2005b)     178  Figure A.1 Department of Fisheries research trawl locations in the Andaman Sea and Gulf of Thailand.   179  Figure A.2 Comparison of annual catch rates scaled from multiple time periods among fishers in different regions in Thailand. Seahorse catch from trawlers and gillnets are shown in (a) and (b) respectively. Southern Gulf of Thailand (S GoT) includes surveyed provinces Chumphon, Surat Thani, and Nakon Si Thammarat, Central and East Gulf of Thailand (C&E GoT) includes Samut Sakhon and Trat, and Andaman Coast (Andaman) includes Krabi, Trang, Phang-nga, Phuket, and Satun.  180 Appendix B  Supporting material for Chapter 3 B.1 Methods estimating number of traders in each trade level With mean trade volume per trader obtained in our interviews, we used annual export volume (top-down) and seahorse catch estimate (bottom-up) to estimate the number of people involved in dried seahorse trade in Thailand. Because lack of data for trade volume from level 3 traders, we only conducted the estimates for level 1, 2, 4, and 5. We used otter trawlers as the representative gear since they caught the majority of seahorses in Thailand (Aylesworth et al. in review). For each way, we applied two methods: Method 1 considering the reported sell and purchase volume separately, and Method 2 considering either sell volume or purchase volume, or the larger of the two reported volumes if a trader reported both.   Method 1: Reported selling and purchasing volumes were considered separately  The calculations of number of traders are as below. All volumes used in the calculations are reported in Table 2. (1) Top-down (illustrated in Figure B1): Number of exporters (level 5) = total export volume / mean ‘sell volume’ per exporter Number of consolidators (level 4) = [(Number of exporters ×mean ‘purchase volume’ per exporter) + total domestic trade volume] / mean ‘sell volume’ per consolidator Number of primary buyers (level 2) = (Number of consolidators × mean ‘purchase volume’ per consolidator) / mean ‘sell volume’ per primary buyer Number of trawlers (level 1) = (Number of primary buyers × mean ‘purchase volume’ per primary buyer) / mean ‘sell volume’ per trawler.    181  Figure B.1 Schematic illustrating the top-down calculations for number of traders in each trade level.   (2) Bottom-up:   182 When calculating from the overall catch volume, we assumed that all catch is sold into trade chain.  Number of trawlers (level 1) = Known (Thailand Department of Fisheries 2014)  Number of primary buyers (level 2) = Total catch / mean ‘purchase volume’ per primary buyer  Number of consolidators (level 4) = (Number of primary buyers × mean ‘sell volume’ per primary buyers) / mean ‘purchase volume’ per consolidator  Number of exporters (level 5) = [(Number of consolidators × mean ‘sell volume’ per consolidator) – total domestic volume] / mean ‘purchase volume’ per exporter  Method 2: Trade volume, the larger one of reported sell or purchase volume  (1) Top-down:  Number of traders in each level = Total export volume / mean ‘trade volume’ per trader in that level  (2) Bottom-up: Number of traders in each level = Total catch volume / mean ‘trade volume’ per trader in that level    183 B.2 Segmented regression for examining the effects of CITES interventions on the seahorse trade volume and prices reported by Thai traders  To examine the effects of the three interventions (CITES listing, CITES implementation, and Thai export quota), we conducted an intervention analysis on the historical volumes and prices reported in the interviews. In the analysis, we investigate the changes from the following year of the three events (2003, 2005, and 2013). For each trade level, we regressed the volume/prices with time, and considered a dummy variable for the intervention (before intervention=0, after intervention=1). The model was as below:   𝑌𝑡 = 𝛽0 + 𝛽1𝑡 + 𝛽2𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛𝑡 + 𝛽3𝑡 ∗ 𝐼𝑛𝑡𝑒𝑟𝑣𝑒𝑛𝑡𝑖𝑜𝑛𝑡   (Eq. 3B.1) Here Yt is the reported trade volume or price at year t. In this model, 𝛽1 describes the mean change in the trade volume/price with time (slope of the regression), 𝛽2 represents the level change in mean volume/price right after the intervention (changes in intercept), and 𝛽3 shows the changes in slope after considering the data after intervention. We only conducted the analysis when data were allowed (more than 4 data points in both time periods: before and after each intervention).  Our intervention analysis showed that the interventions did not coincide with the changes in trade volume; however, they might correspond with the increases in selling prices. There was no apparent effect of the implementation of the voluntary quota on trade volumes for traders in level 2-5, nor for CITES implementation for primary buyers (level 2, Table B.1). Data were not available to look at the possible effect of the CITES listing and implementation for exporters and consolidators; also too few trade volume before 2003 (n<4) was reported to conduct analysis for   184 the effect of the CITES listing for primary buyers. In contrast to the trade volume, the intervention analysis showed that selling prices increased significantly after 2013 for fishers (level 1), primary buyers (level 2), and consolidators (level 4) (Table B.1). The increasing rate of selling prices became higher after CITES listing and also after CITES implementation for fishers and primary buyers (Table B.1). We had too few data (n<4) for historic prices reported by primary buyers for years before 2003, and by consolidators and exporters for years before 2005 to examine the effects of the two interventions for traders in these levels.  185 Table B.1 Results of intervention analysis for the changes in trade volume and selling prices corresponding with the three events: CITES listing, CITES implementation, and national quota setting. Estimates of regressions are shown with the p-value in brackets. Significant results (p<0.05) are in bold and italicized. If there were fewer than four data points for a time period (before or after an intervention), no analysis was applied.   CITES listing CITES implementation Thai voluntary export quota   Volume Price Volume Price Volume Price Exporters (level 5) Trend (1) -- -- -- -- 0.06 (0.65) 0.20 (0.10) Changes in intercept (2) -- -- -- -- 238.23 (0.85) -1054.04 (0.21) Changes in slope  (3) -- -- -- -- -0.12 (0.85) 0.52 (0.21) Consolidators (level 4) Trend (1) -- -- - -- -0.12 (0.23) -0.12 (0.19) Changes in intercept (2) -- -- - -- -1842.00 (0.20) 3.69 (<0.01) Changes in slope  (3) -- -- -- -- 0.91 (0.20) NA Primary buyers (level 2) Trend (1) -- 0.05 (0.26) 0.04 (0.83) 0.05 (0.26) -0.31 (0.02) -0.01 (0.71) Changes in intercept (2) -- -0.00 (<0.01) -1223 (0.16) -0.00 (<0.01) 2.67 (0.08) 2.83 (<0.01) Changes in slope  (3) -- 0.98 (<0.01) 0.61 (0.16) 0.98 (<0.01) NA NA Fishers (level 1) Trend (1) -- -0.02 (0.51) -- -0.02 (0.51) -- -0.04 (0.09) Changes in intercept (2) -- -1,398 (<0.01) -- -0.00 (<0.01) -- 2.85 (<0.01)   186 Appendix C  Supporting material for Chapter 4 C.1 The sources of the country attributes 1. Demersal fish catch: The estimates of annual demersal fish catch for each country from 1983-2013 were provided by the Sea Around Us include commercial and artisanal fisheries, all demersal fishes, downloaded from the Sea Around Us database (http://www.seaaroundus.org/) at January 16th, 2017.  2. Gross domestic product (GDP) per capita data were obtained from World Bank at July 11th, 2016, except for Taiwan, where the data was not included. Taiwan’s data was retrieved from the International Monetary Fund (IMF) database (http://www.econstats.com/weo/V001.htm) at July 11th, 2016. All GDP data were retrieved for 1983-2014. 3. Number of fishers in the marine sector, including commercial and small-scale fishers, was obtained from (Teh and Sumaila 2013).  4. Distances data were procured from the CEPII database (http://www.cepii.fr/CEPII/en/bdd_modele/presentation.asp?id=6) at June 22nd, 2016. The calculations for the bilateral distances weighted the populations contribution of the top 25 cities in each country (Head and Mayer 2010).  5. Trade values in general goods with China of each country were collected from the China Statistical Yearbook, from 2005-2013 (http://www.stats.gov.cn/tjsj/ndsj/) at June 21st, 2016.    187 C.2 The results of iterative segmented regressions. Table C.1 Results of testing the effects of CITES interventions on Hong Kong’s import volume, using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂𝟏𝒀𝒆𝒂𝒓𝒕 + 𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰).  Breakpoint: 2003 r2= 0.86 Estimate P-value Breakpoint: 2006 r2= 0.67 Estimate P-value Intercept -8435427 <0.01 Intercept -1524316 0.36 Year, 𝑎1 4228 <0.01 Year, 𝑎1 771 0.36 Slope change, 𝑎2 -6281 <0.01 Slope change, 𝑎2 -2214 0.06 Intercept change, 𝑎3 12569896 <0.01 Intercept change, 𝑎3 4430931 0.06 Breakpoint: 2004 r2= 0.86   Breakpoint: 2007 r2= 0.59   Intercept -6285772 <0.01 Intercept 31104 0.98 Year, 𝑎1 3152 <0.01 Year, 𝑎1 -5 0.99 Slope change, 𝑎2 -4990 <0.01 Slope change, 𝑎2 -1445 0.24 Intercept change, 𝑎3 9987198 <0.01 Intercept change, 𝑎3 2892619 0.24 Breakpoint: 2005 r2= 0.92    Intercept -5023697 <0.01 Year, 𝑎1 2521 <0.01 Slope change, 𝑎2 -3807 <0.01 Intercept change, 𝑎3 7614557 <0.01   188 Table C.2 Results of testing the effects of CITES interventions on Taiwan’s import volume, using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂𝟏𝒀𝒆𝒂𝒓𝒕 + 𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰). Breakpoint: 2000 r2= 0.77 Estimate P-value Breakpoint: 2004 r2= 0.74 Estimate P-value Intercept -601837 <0.01 Intercept -357885 0.02 Year, 𝑎1 306 <0.01 Year, 𝑎1 184 0.02 Slope change, 𝑎2 -1077 <0.01 Slope change, 𝑎2 -837 <0.01 Intercept change, 𝑎3 2152814 <0.01 Intercept change, 𝑎3 1673521 <0.01 Breakpoint: 2001 r2= 0.79   Breakpoint: 2005 r2= 0.66   Intercept -634836 <0.01 Intercept -283419 0.05 Year, 𝑎1 323 <0.01 Year, 𝑎1 146 0.04 Slope change, 𝑎2 -1018 <0.01 Slope change, 𝑎2 -708 0.01 Intercept change, 𝑎3 2033705 <0.01 Intercept change, 𝑎3 1413147 0.01 Breakpoint: 2002 r2= 0.76    Breakpoint: 2006 r2= 0.69   Intercept -497792 <0.01 Intercept  -163435 0.26 Year, 𝑎1 254 <0.01 Year, 𝑎1  86 0.24 Slope change, 𝑎2 -990 <0.01 Slope change, 𝑎2  -663 0.04 Intercept change, 𝑎3 1979482 <0.01 Intercept change, 𝑎3  1324317 0.04 Breakpoint: 2003  r2= 0.74   Breakpoint: 2007 r2= 0.72   Intercept -423618 <0.01 Intercept  -121370 0.36 Year, 𝑎1 217 0.01 Year, 𝑎1  65 0.33   189     Breakpoint: 2003  r2= 0.74 Estimate P-value    Breakpoint: 2007 r2= 0.72 Estimate P-value Slope change, 𝑎2 -932 <0.01 Slope change, 𝑎2  -330 0.35 Intercept change, 𝑎3 1863985 <0.01 Intercept change, 𝑎3  655508 0.35   190 Table C.3 Results of testing the effects of CITES interventions on the number of source countries of Hong Kong, using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂 𝒀𝒆𝒂𝒓𝒕 + 𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰).   Breakpoint: 2003 r2= 0.77 Estimate P-value Breakpoint: 2006 r2= 0.82 Estimate P-value Intercept -1987 <0.01 Intercept 132 0.84 Year, 𝑎1 1 <0.01 Year, 𝑎1 -0.06 0.86 Slope change, 𝑎2 -1.7797 <0.01 Slope change, 𝑎2 -0.14 0.74 Intercept change, 𝑎3 3560 <0.01 Intercept change, 𝑎3 275 0.75 Breakpoint: 2004 r2= 0.78   Breakpoint: 2007 r2= 0.74   Intercept -1130 0.32 Intercept 980 0.15 Year, 𝑎1 1 0.32 Year, 𝑎1 -0.48 0.16 Slope change, 𝑎2 -1 0.05 Slope change, 𝑎2 0.30 0.55 Intercept change, 𝑎3 2488 0.05 Intercept change, 𝑎3 -616 0.55 Breakpoint: 2005 r2= 0.92    Intercept -1130 0.05 Year, 𝑎1 1 0.05 Slope change, 𝑎2 -1 0.01 Intercept change, 𝑎3 1757 0.01   191 Table C.4 Results of testing the effects of CITES interventions on the number of source countries of Taiwan, using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂𝟏𝒀𝒆𝒂𝒓𝒕 + 𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰). Breakpoint: 2000 r2= 0.77 Estimate P-value Breakpoint: 2004 r2= 0.77 Estimate P-value Intercept 104 0.38 Intercept 89 0.30 Year, 𝑎1 -0.05 0.41 Year, 𝑎1 -0.04 0.34 Slope change, 𝑎2 -0.40 <0.01 Slope change, 𝑎2 -0.33 0.01 Intercept change, 𝑎3 795 <0.01 Intercept change, 𝑎3 662 0.01 Breakpoint: 2001 r2= 0.77   Breakpoint: 2005 r2= 0.80   Intercept 99 0.37 Intercept 84 0.27 Year, 𝑎1 -0.05 0.40 Year, 𝑎1 -0.04 0.30 Slope change, 𝑎2 -0.41 <0.01 Slope change, 𝑎2 -0.21 0.11 Intercept change, 𝑎3 826 <0.01 Intercept change, 𝑎3 417 0.11 Breakpoint: 2002 r2= 0.77    Breakpoint: 2006 r2= 0.74   Intercept 94 0.35 Intercept  166 0.05 Year, 𝑎1 -0.04 0.39 Year, 𝑎1  -0.08 0.05 Slope change, 𝑎2 -0.42 <0.01 Slope change, 𝑎2  -0.25 0.14 Intercept change, 𝑎3 836 <0.01 Intercept change, 𝑎3  506 0.14 Breakpoint: 2003  r2= 0.79   Breakpoint: 2007 r2= 0.74   Intercept 32 0.72 Intercept  187 0.02 Year, 𝑎1 -0.01 0.77 Year, 𝑎1  -0.09 0.02   192  Breakpoint: 2003  r2= 0.79 Estimate P-value Breakpoint: 2007 r2= 0.74 Estimate P-value Slope change, 𝑎2 -0.34 <0.01 Slope change, 𝑎2  -0.16 0.43 Intercept change, 𝑎3 680 <0.01 Intercept change, 𝑎3  317 0.43   193 Table C.5 Results of testing the effects of CITES interventions on the evenness of supply to Hong Kong, using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂𝟏𝒀𝒆𝒂𝒓𝒕 + 𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰).    Breakpoint: 2003 r2= 0.74 Estimate P-value Breakpoint: 2006 r2= 0.82 Estimate P-value Intercept -12.32 0.19 Intercept -8.03 0.05 Year, 𝑎1 0.01 0.16 Year, 𝑎1 0.00 0.03 Slope change, 𝑎2 0.00 0.54 Slope change, 𝑎2 0.00 0.14 Intercept change, 𝑎3 5.82 0.54 Intercept change, 𝑎3 7.53 0.14 Breakpoint: 2004 r2= 0.77   Breakpoint: 2007 r2= 0.81   Intercept -6.98 0.30 Intercept -9.94 0.01 Year, 𝑎1 0.00 0.24 Year, 𝑎1 0.01 0.00 Slope change, 𝑎2 0.00 0.75 Slope change, 𝑎2 -0.01 0.05 Intercept change, 𝑎3 2.31 0.75 Intercept change, 𝑎3 10.44 0.05 Breakpoint: 2005 r2= 0.82    Intercept -5.87 0.22 Year, 𝑎1 0.00 0.16 Slope change, 𝑎2 0.00 0.49 Intercept change, 𝑎3 3.71 0.49   194 Table C.6 Results of testing the effects of CITES interventions on the evenness of supply to Taiwan, using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂𝟏𝒀𝒆𝒂𝒓𝒕 + 𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰). Breakpoint: 2000 r2= 0.58 Estimate P-value Breakpoint: 2004 r2= 0.44 Estimate P-value Intercept -7.57 0.01 Intercept -0.98 0.65 Year, 𝑎1 0.00 0.00 Year, 𝑎1 0.00 0.39 Slope change, 𝑎2 0.00 0.17 Slope change, 𝑎2 0.00 0.16 Intercept change, 𝑎3 -5.67 0.17 Intercept change, 𝑎3 -8.74 0.16 Breakpoint: 2001 r2= 0.58   Breakpoint: 2005 r2= 0.44   Intercept -6.34 0.01 Intercept -1.00 0.62 Year, 𝑎1 0.00 0.00 Year, 𝑎1 0.00 0.35 Slope change, 𝑎2 0.00 0.04 Slope change, 𝑎2 0.00 0.28 Intercept change, 𝑎3 -8.81 0.04 Intercept change, 𝑎3 -7.50 0.28 Breakpoint: 2002 r2= 0.63    Breakpoint: 2006 r2= 0.43   Intercept -5.74 0.01 Intercept  -1.99 0.30 Year, 𝑎1 0.00 0.00 Year, 𝑎1  0.00 0.14 Slope change, 𝑎2 0.01 0.00 Slope change, 𝑎2  0.01 0.18 Intercept change, 𝑎3 -12.69 0.00 Intercept change, 𝑎3  -11.13 0.18 Breakpoint: 2003  r2= 0.45   Breakpoint: 2007 r2= 0.44   Intercept -2.30 0.32 Intercept  -1.50 0.40   195   Breakpoint: 2003  r2= 0.45 Estimate P-value Breakpoint: 2007 r2= 0.44 Estimate P-value Year, 𝑎1 0.00 0.17 Year, 𝑎1  0.00 0.19 Slope change, 𝑎2 0.01 0.05 Slope change, 𝑎2  0.00 0.53 Intercept change, 𝑎3 -11.12 0.05 Intercept change, 𝑎3  -5.98 0.53   196 Table C.7 Results of testing the effects of CITES interventions on Hong Kong’s import prices, using iterative segmented regressions (𝑽𝒕 = 𝒂𝟎 + 𝒂𝟏𝒀𝒆𝒂𝒓𝒕 + 𝒂𝟐𝒀𝒆𝒂𝒓𝒕 ∙ 𝑰 + 𝒂𝟑𝑰).  Breakpoint: 2003 r2= 0.60 Estimate P-value Breakpoint: 2006 r2= 0.61 Estimate P-value Intercept 216 0.56 Intercept 61 0.73 Year, 𝑎1 -0.11 0.57 Year, 𝑎1 -0.03 0.75 Slope change, 𝑎2 0.29 0.15 Slope change, 𝑎2 0.25 0.04 Intercept change, 𝑎3 -577 0.15 Intercept change, 𝑎3 -510 0.04 Breakpoint: 2004 r2= 0.61   Breakpoint: 2007 r2= 0.61   Intercept 162 0.56 Intercept 12 0.93 Year, 𝑎1 -0.08 0.57 Year, 𝑎1 0.00 0.96 Slope change, 𝑎2 0.27 0.08 Slope change, 𝑎2 0.26 0.04 Intercept change, 𝑎3 -548 0.08 Intercept change, 𝑎3 -516 0.04 Breakpoint: 2005 r2= 0.61    Intercept 111 0.61 Year, 𝑎1 -0.05 0.62 Slope change, 𝑎2 0.26 0.05 Intercept change, 𝑎3 -524 0.05   197 Appendix D  Supporting material for Chapter 5 Table D.1 The ISO-2 code of countries and territories that were mentioned in the main text. ISO code Full name AU Australia BJ Benin BR Brazil CA Canada CD Democratic Republic of the Congo CH Switzerland CM Cameroon CN China CZ Czech Republic DE Germany DK Denmark ES Spain FJ Fiji FR France GB United Kingdom of Great Britain and Northern Ireland GL Greenland GN Guinea GY Guyana HK Hong Kong ID Indonesia IT Italy JP Japan   198 ISO code Full name KR Republic of Korea ML Mali MY Malaysia NL Netherlands PE Peru RU Russian Federation SB Solomon Islands SG Singapore SR Suriname TG Togo TH Thailand TO Tonga TZ United Republic of Tanzania US United States of America UZ Uzbekistan VN Viet Nam ZA South Africa ZW Zimbabwe    199 Table D.2 The list of CITES Appendix II marine species Common name Order Family Genus Species Notes All antipatharian (black coral) species Antipatharia    Coral All stony coral species Scleractinia    Coral Blue coral    Heliopora coerulea Coral Organ-pipe Coral    Tubipora musica Coral Stylasteridae  Stylasteridae   Coral Fire Coral  Milleporidae   Coral All giant clam species  Tridacnidae    Basking shark    Cetorhinus maximus  Great White Shark    Carcharodon carcharias  Scalloped hammerhead Sharks    Sphyrna lewini  Smooth hammerhead Sharks     Sphyrna zygaena  Great hammerhead Sharks     Sphyrna mokarran  Humphead wrasse    Cheilinus undulatus  Manta Rays   Manta   Sturgeons   Acipenseriformes  Oceanic Whitetip Shark    Carcharhinus longimanus  Porbeagle Shark    Lamna nasus    200 Common name Order Family Genus Species Notes Queen conch    Lobatus gigas  All seahorse species   Hippocampus  Whale shark    Rhincodon typus  European eel    Anguilla anguilla  Date shell    Lithophaga lithophaga      201 Table D.3 The results of linear regression of species richness in exports versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24).         Country/territory Year Intercept Coefficient SE p-value Coefficient SE p-value ID 0.008 0.001 <0.01 -16.10 2.553 <0.01 AU 0.011 0.002 <0.01 -21.25 3.051 <0.01 US 0.006 0.001 <0.01 -11.66 2.695 <0.01 FJ 0.003 0.001 <0.01 -6.34 2.121 <0.01 SB 0.003 0.001 <0.01 -6.18 1.003 <0.01 TO 0.005 0.001 <0.01 -9.47 1.763 <0.01 MG -0.002 0.001 <0.01 3.53 1.085 <0.01 GY 0.000 0.000 0.382 0.77 0.799 0.346 SR -0.002 0.000 <0.01 3.45 0.417 <0.01 CA 0.000 0.000 0.931 -0.01 0.609 0.982   202 Table D.4 The results of linear regression results of species richness in imports versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24).         Country/territory Year Intercept Coefficient SE p-value Coefficient SE p-value US 0.007 0.002 <0.01 -12.665 3.532 <0.01 JP -0.003 0.002 0.161 6.064 3.858 0.130 HK 0.010 0.001 <0.01 -20.250 1.574 <0.01 DE 0.004 0.002 0.036 -8.419 3.907 0.042 KR 0.010 0.001 <0.01 -18.980 1.417 <0.01 GB -0.003 0.002 0.089 6.244 3.352 0.076 CA 0.009 0.002 <0.01 -18.126 3.779 <0.01 FR 0.002 0.002 0.146 -4.478 3.140 0.168 NL -0.007 0.001 <0.01 14.903 1.760 <0.01 CN 0.008 0.001 <0.01 -15.660 1.681 <0.01          203 Table D.5 The results of linear regression of out-degree versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24)Country/territory Year Intercept Coefficient SE p-value Coefficient SE p-value US 3.152 0.207 <0.01 -6247.812 414.735 <0.01 ID 2.137 0.207 <0.01 -4229.464 413.814 <0.01 ZA 1.017 0.348 <0.01 -2007.418 696.315 <0.01 VN 1.042 0.088 <0.01 -2070.000 177.100 <0.01 CA 0.378 0.125 <0.01 -723.884 250.939 <0.01 GY 1.094 0.153 <0.01 -2166.815 307.244 <0.01 TZ 0.112 0.246 0.654 -195.966 493.027 0.695 BR 0.967 0.147 <0.01 -1927.835 295.076 <0.01 MY 0.283 0.130 0.040 -545.836 260.101 0.048 FJ 1.115 0.074 <0.01 -2221.000 148.000 <0.01   204 Table D.6 The results of linear regression of in-degree versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24).         Country/territory Year Intercept Coefficient SE p-value Coefficient SE p-value US -0.445 0.277 0.122 966.631 554.501 0.095 JP 0.611 0.185 <0.01 -1177.808 369.807 <0.01 HK 1.282 0.121 <0.01 -2543.599 242.347 <0.01 DE 0.947 0.259 <0.01 -1858.359 518.441 <0.01 KR 1.213 0.084 <0.01 -2412.000 167.800 <0.01 IT 0.308 0.124 0.021 -586.001 248.524 0.028 CN 1.399 0.071 <0.01 -2783.000 143.100 <0.01 FR 0.251 0.232 0.291 -467.820 464.832 0.325 GB 0.293 0.152 0.067 -555.407 304.112 0.081 SG 0.550 0.126 <0.01 -1079.750 252.971 <0.01   205 Table D.7 The results of linear regression of the ranking of closeness centrality versus year for the 2014 top 10 countries/territories. Data were from 1991-2014 (n=24).             Country/territory Year Intercept Coefficient SE p-value Coefficient SE p-value US <0.001 <0.001 0.0973 1.000 <0.001 <0.01 ID -0.338 0.078 <0.01 681.038 156.595 <0.01 CA 0.033 0.029 0.261 -64.165 58.153 0.282 ZA -0.272 0.088 <0.01 550.862 176.320 <0.01 JP 0.007 0.043 0.872 -9.514 85.181 0.912 DE -0.203 0.077 0.015 412.807 153.483 0.013 FR 0.019 0.073 0.795 -30.809 145.507 0.834 HK -0.186 0.047 <0.01 386.722 94.222 <0.01 MY 0.074 0.090 0.420 -137.344 180.279 0.454 CN -0.113 0.080 0.170 240.282 160.314 0.148 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0355398/manifest

Comment

Related Items